Download - Proefschrift Waterschoot

Nice to have or need to have?

Unraveling Dosage of Pain Rehabilitation

Franka Waterschoot

The research project “Nice to have or need to have” was funded by:

Pijnrevalidatie - Centrum voor revalidatie UMCG

The publication of this thesis was fi nancially supported by:

University of Groningen - RUG

Graduate school for Health Research - SHARE

Centrum voor Revalidatie - Universitair Medisch Centrum Groningen (UMCG)

Pijnrevalidatie - Centrum voor revalidatie UMCG

Stichting Beatrixoord Noord - Nederland

Menzis Zorgverzekering

Cover design: Isontwerp - ‘s Hertogenbosch

Layout: Gildeprint - Enschede

Printed by: Gildeprint - Enschede

Waterschoot, Franka. Nice to have or need to have. Unraveling dosage of pain rehabilitation.

Thesis University of Groningen, the Netherlands – with references – with summary in Dutch.

ISBN: 978-90-367-8512-9 (printed version)

ISBN: 978-90-367-8513-6 (electronic version)

Copyright © 2015. F.P.C. Waterschoot.

Nice to have or need to have?

Unraveling Dosage of Pain Rehabilitation

Proefschrift

ter verkrijging van de graad van doctor aan de

Rijksuniversiteit Groningen

op gezag van de

rector magnifi cus prof. dr. E. Sterken

en volgens besluit van het College voor Promoties.

De openbare verdediging zal plaatsvinden op

maandag 15 februari 2016 om 14.30 uur

door

Franka Petronella Catharina Waterschoot geboren op 20 juni 1978

te Hoogeloon

Promotores

Prof. dr. M.F. Reneman

Prof. dr. J.H.B. Geertzen

Prof. dr. P.U. Dijkstra

Beoordelingscommissie

Prof. dr. J.G.M. Rosmalen

Prof. dr. K.M.G. Schreurs

Prof. dr. C. van Bennekom

Paranimfen: Annemieke de Jong

Berry van Holland

Table of content

Chapter 1 General introduction. 9

Chapter 2 Dose or content? Effectiveness of pain rehabilitation programs for 19

patients with chronic low back pain: A systematic review.

Chapter 3 Course of disability reduction during a pain rehabilitation program: 43

A prospective clinical study.

Chapter 4 Dosage of pain rehabilitation programs for patients with chronic 57

musculoskeletal pain: A non- inferiority randomized controlled trial.

Chapter 5 Dosage of pain rehabilitation programs: A qualitative study from 73

patient and professionals’ perspectives.

Chapter 6 Towards a case complexity index of patients with chronic 91

non-specifi c musculoskeletal pain: A Delphi and feasibility study.

Chapter 7 General discussion. 109

Summary 123

Samenvatting 129

Dankwoord 137

Over de auteur 143

Research institute SHARE 147

Extremities, Pain and Disability (Expand) 153

R1R2R3R4R5R6R7R8R9

R10R11R12R13R14R15R16R17R18R19R20R21R22R23R24R25R26R27R28R29R30R31R32R33R34R35R36R37R38R39

10 | Chapter 1

R1R2R3R4R5R6R7R8R9R10R11R12R13R14R15R16R17R18R19R20R21R22R23R24R25R26R27R28R29R30R31R32R33R34R35R36R37R38R39

General introduction | 11

1Background of thesis

In Europe, 19% of the adult population suffers from chronic pain of moderate to severe intensity 1. Chronic pain is associated with lower quality of life, less participation in major life domains,

higher healthcare utilization, lower work productivity, and higher direct and indirect costs 1,2.

Among the many treatment options for chronic pain are multidisciplinary pain rehabilitation

programs (PRPs). There is robust evidence that PRPs are effective 3-5. They can help to improve

quality of life, improve participation in major life domains, and decrease healthcare utilization.

There is confl icting evidence on the benefi ts for work-related outcomes, however 5,6.

The positive effects are ro bust, although the average size of the effects is modest and there is a

great need for further improvements. Optimizing the outcomes of PRPs should benefi t patients

suffering from chronic pain, and the benefi ts should extend to the patients’ families, in addition

to other stakeholders such as rehabilitation centers, healthcare providers, healthcare insurers,

employers, coworkers and disability insurers (temporary and permanent). Up to now, all studies

that have aimed at optimizing outcomes of PRP have focused on content, composition and

delivery form, such as individually or multidisciplinary therapy 3-10. However, optimal PRP may

also relate to dosage of PRP. Dosage of PRP appears to be a neglected, but potentially relevant,

topic in PRP and PRP research. Dose variables can be expressed as the total duration and amount

of contact hours (Text box). Additionally, the number of healthcare professionals involved in the

program, regardless of whether an inpatient or an outpatient program is offered, and dosage per

contact hour are related to the total dosage of PRP.

There are several motives for arguing why a closer look at the dosage of PRP might be necessary:

- In clinical practice, large variations of PRP-dosage are observed, in the Netherlands as well as

internationally. There are also large variations observed in dosages that have been described

in PRP research 6,7. The rationale for dosage variations is unknown.

- Scientifi cally, a consequence of research focused on content only is that infl uence of

dosage on the effect of PRP is neglected. For example, a study concluding inferiority of PRP,

comparing multidisciplinary PRP of 100 hours with mono-disciplinary physiotherapy of 50

hours could be confounded because not only does the content differ, but also dosage does

differ. Theoretically, both content and dose variables can contribute to the effectiveness

of PRP. There is a lack of studies comparing multidisciplinary PRPs with only dosage as a

comparing variable: randomized controlled trials comparing differences in dosage between

multidisciplinary treatments with the same content have not yet been performed.

- Viewed from the process of care, optimizing the entire logistic process of PRP, from registration

to discharge, is important. It can reduce waiting lists and improve cost effectiveness,

benefi tting both patients and rehabilitation centers. To manage this process, it is necessary

R1R2R3R4R5R6R7R8R9


12 | Chapter 1

to determine the total number of patients treated, as well as the in- and outfl ow of patients.

Consequently, to infl uence the outfl ow to optimize infl ow, it is necessary to determine the

offered dosage of PRP.

- Ethically, not knowing the optimal dosage of PRP, the question is raised as to whether over-

or undertreatment will occur.

- Developments in the Dutch healthcare system require more transparency regarding quality

and fi nancial sustainability of specialist medical care. This transparency will lead to a better

understanding of rationale of care, which will benefi t patients, health insurers and healthcare

providers. These developments call for comparability of treatments with regard to content,

composition, dosage, and fi nances.

Therefore, the time has come to include dose variables in the question of “what works for whom”,

and therefore to analyze dosage at the level of the individual patient related to accomplishing

their goals, and analyzing dosage related to effect from the perspective of different stakeholders

to optimize the cost-effectiveness of multidisciplinary PRPs.

Textbox Defi nitions related to dosage of pain rehabilitation programs

Defi nitionsPain rehabilitation programs Multidisciplinary treatments for patients suffering from chronic

musculoskeletal pain, existing of a combination of physical and psychosocial components, with fi nal responsibility of the rehabilitation physician to coach patients on managing their pain and understanding the consequences of pain in daily life 11

Chronic pain Pain without apparent biological value that has persisted beyond the normal tissue healing time (usually taken to be three months) 12

Disability Disability is an umbrella term for impairments, activity limitations and participation restrictions referring to the negative aspects of the interaction between a person’s health condition(s) and that individual’s contextual factors (environmental and personal factors) 13

Dosage variables:• Duration of program Total length of program expressed in weeks or months, from fi rst to last

contact

• Intensity of program The number of contact hours per week or month• Intensity of content/exercise The extent of exercise or other content per unit of time, frequency of

exercises per session, intensity of the forces applied or exertion required

• Contact hours Total number of hours that a patient spends with his rehabilitation professional during the program

• Number of healthcare professionals

The number of different disciplines (psychologists, physiotherapists, occupational therapists) involved in the program

Generally, patients with chronic pain improve signifi cantly in several domains (pain-related

disability, return to work, quality of life) after PRP 5,6. But it is presently unknown how much

treatment is needed to achieve the benefi cial effects or enhance the effect size. Multidisciplinary



1PRPs are based on the bio-psycho-social model that aims to reduce pain-related disability. The

cognitive-behavioral components assume that patients change their pain management skills

during PRP. These behavioral changes support the reduction of pain-related disability without the

aim of pain reduction. Consequently, aiming for behavioral changes during PRP, the dosage of

PRP could be based on the dosage of other therapies that use cognitive-behavioral therapy for

chronic pain. Unfortunately, the optimum dosage of those cognitive-behavioral interventions for

patients with chronic pain is also unknown 14. In conclusion, there is currently a gap in knowledge

regarding dosage of PRP based on cognitive-behavioral principles.

In order to fi nd the optimum dosage of PRP, underlying constructs of dosage of PRP should be

explored. When unraveling this topic in pain rehabilitation research, multiple questions were

raised. Some questions are as follows:

• What is the rationale for the dosage of the present programs in today’s daily practice in

Dutch PRP?

• Is therapy needed each week or is a time-out also possible?

• Which healthcare professionals should be involved in PRP and what is their added value in

PRP?

• Which patients require higher dosage of PRP and for whom would a lower dosage be

suffi cient?

• What are reasons to stop PRP and how do these reasons relate to the desired improvements?

• How does reduction of disability occur during weeks of PRP?

• Can the dosage of PRP be reduced without a reduction in the effects?

• Should there be different dosages for patients that exhibit differences in case complexity?

Aim and research questions

The overall aim of the thesis is to acquire insight into the effects of dosage of multidisciplinary

pain rehabilitation programs for patients with chronic pain.

This overall aim is specifi ed in the following research questions:

• What is the reported dosage of treatment for patients with chronic low back pain and what

is the infl uence of dose variables on the outcome of PRPs for patients with chronic low back

pain? (Chapter 2).

• Which course does disability reduction during PRP follow and which factors infl uence this

course over time? (Chapter 3).

• Is a shorter form of PRP non-inferior to the PRP in care as usual? Do effects differ signifi cantly?

(Chapter 4).

R1R2R3R4R5R6R7R8R9


14 | Chapter 1

• What are the differences and similarities in the experiences and perspectives of patients who

underwent PRP and rehabilitation professionals working in PRPs regarding dosage of PRP

between three centers in the Netherlands? (Chapter 5).

• Which factors determine case complexity of patients with CMP that are eligible for PRP?

(Chapter 6).

Outline of thesis

Since, up to now, dosage has been a neglected topic in pain research, different research methods

and perspectives have been applied in this thesis.

The systematic review and meta-analysis described in Chapter 2 aimed to analyze the infl uence

of dose variables on the outcome of PRPs. There were no studies found that primarily analyzed

the relationship between dose and effects of PRP. Therefore, this review focused on randomized

controlled trials (RCTs) aimed at assessing effectiveness of PRPs that reported dose variables of

the studied programs. The reported dose variables were used to analyze the infl uence of dose

variables on the effects.

The course of disability reduction during PRP and factors infl uencing this course were analyzed

in Chapter 3. The majority of studies analyzed pre- and post-outcomes of disability reduction.

However to acquire insight into the dosage of PRP, it is a prerequisite to analyze the course of the

primary outcome during PRP.

Chapter 4 describes the fi rst randomized controlled trial in PRP research that compared PRPs with

similar content but different dosages. The control arm was PRP performed in care as usual dosage

(8, 12, 16 or 20 weeks) and the experimental arm was PRP performed at a duration four weeks

shorter than the care as usual.

To gain in-depth information on the experiences and perspectives of patients and rehabilitation

professionals regarding dosage of PRP, a qualitative study was performed, which is described in

Chapter 5. Interviews with patients and focus group interviews with rehabilitation professionals

were conducted in three centers in the Netherlands offering PRP with differences in dosage of

PRP, while patient characteristics were similar.

From clinical practice, a relationship was hypothesized between case complexity and the required

dosage of PRP. However, little research is available regarding the interpretation of case complexity

within multidisciplinary PRP. Therefore, it was not possible to test this relationship. In Chapter



16, the operationalization of case complexity was studied using a Delphi technique among

rehabilitation professionals working in pain rehabilitation centers throughout the Netherlands.

In Chapter 7, the general discussion, all fi ndings from the previous chapters are integrated

and discussed. Implications for clinical practice and recommendations for future research are

presented.

R1R2R3R4R5R6R7R8R9


16 | Chapter 1

References

1. Breivik H, Collett B, Ventafridda V, Cohen R, Gallacher D. Survey of chronic pain in Europe: Prevalence, impact on daily life, and treatment. Eur J Pain. 2006;10(4):287-333.

2. Lambeek LC, van Tulder MW, Swinkels IC, Koppes LL, Anema JR, van Mechelen W. The trend in total cost of back pain in the netherlands in the period 2002 to 2007. Spine (Phila Pa 1976). 2011;36(13):1050-1058.

3. Flor H, Fydrich T, Turk DC. Effi cacy of multidisciplinary pain treatment centers: A meta-analytic review. Pain. 1992;49(2):221-230.

4. Scascighini L, Toma V, Dober-Spielmann S, Sprott H. Multidisciplinary treatment for chronic pain: A systematic review of interventions and outcomes. Rheumatology (Oxford). 2008;47(5):670-678.

5. Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ. 2015;350:h444.

6. van Geen JW, Edelaar MJ, Janssen M, van Eijk JT. The long-term effect of multidisciplinary back training: A systematic review. Spine (Phila Pa 1976). 2007;32(2):249-255.

7. Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: Systematic review. BMJ. 2001;322(7301):1511-1516.

8. Smeets RJ, Vlaeyen JW, Hidding A, et al. Active rehabilitation for chronic low back pain: Cognitive-behavioral, physical, or both? fi rst direct post-treatment results from a randomized controlled trial [ISRCTN22714229]. BMC Musculoskelet Disord. 2006;7:5.

9. Dufour N, Thamsborg G, Oefeldt A, Lundsgaard C, Stender S. Treatment of chronic low back pain: A randomized, clinical trial comparing group-based multidisciplinary biopsychosocial rehabilitation and intensive individual therapist-assisted back muscle strengthening exercises. Spine (Phila Pa 1976). 2010;35(5):469-476.

10. Kainz B, Gulich M, Engel EM, Jackel WH. Comparison of three outpatient therapy forms for treatment of chronic low back pain-- fi ndings of a multicentre, cluster randomized study. Rehabilitation (Stuttg). 2006;45(2):65-77.

11. Köke A, Brouwers M, Heuts P, et al. Consensus rapport pijnrevalidatie Nederland. 2005.

12. International Association for the Study of Pain. Task Force on Taxonomy. Classifi cation of chronic pain: Descriptions of chronic pain syndromes and defi nitions of pain terms. 2nd ed. Seattle: IASP Press; 1994.

13. World HO, Organisation Mondiale de lS, World HO. How to use the ICF: A practical manual for using the international classifi cation of functioning, disability and health (ICF): Exposure draft for comment. Genf: WHO; 2013.

14. Ehde DM, Dillworth TM, Turner JA. Cognitive-behavioral therapy for individuals with chronic pain: Effi cacy, innovations, and directions for research. Am Psychol. 2014;69(2):153-166.



1

FrFrFrannankakaka PP.C.CC.. WaWaW tetet rsrsrschchchoooooot,t,t, PPPieeeteteter r U.U.U. DDDijijijkskskstrtrraaa PhPhP D,D,D, NNNieieek kk HoHoHolll akakak MMMScScSc, HaHaHaititi zezeze JJJ.. dedede VVVrirr esess PPPhDhDhD, Jaannn H.H.H.B.BB. GGGeeeee rtrtr zezez nn MDMDD,, PhPhPhD,D,D, MMMiccchihhielele FFF. ReReReneneemamaannn PhPhP D.D

PaPaP ininin, 20202014141 ,1,11555555 (((1)1)1):: 1717179-9--898989..

R1R2R3R4R5R6R7R8R9


20 | Chapter 2

Abstract

We sought to systematically analyze the infl uence of dose of Pain Rehabilitation Programs (PRPs)

for patients with chronic low back Pain (CLBP) on disability, work participation and quality of life

(QoL). Literature searches were performed in PUBMED, Cochrane library, Cinahl and EMBASE up

to October 2012, using MeSH terms, other relevant terms and free-text words.

Randomized controlled trials in English, Dutch and German, analyzing the effect of PRPs, were

included. One of the analyzed interventions had to be a PRP. Outcomes should be reported

regarding disability, work participation or QoL. To analyze dose, the number of contact hours

should be reported. Two reviewers independently selected titles, abstract and full text articles on

the basis of inclusion and exclusion criteria. Data were extracted and risk of bias was assessed. Effect

sizes (ES) were calculated for each intervention, and infl uence of dose variables was analyzed by a

mixed model analysis. Eighteen studies were identifi ed, reporting a wide variety of dose variables

and contents of PRPs. Analyses showed that evaluation moment, number of disciplines, type of

intervention, duration of intervention in weeks, percentage of women, and age, infl uenced the

outcomes of PRPs. The independent effect of dose variables could not be distinguished from

content because these variables were strongly associated. Because dose variables were never

studied separately or reported independently, we were not able to disentangle the relationship

between dose, content, and effects of PRPs on disability, work participation and QoL.

Keywords

Chronic low back pain, dose, effectiveness, pain rehabilitation programs


Dose or content? Effectiveness of PRP | 21

2

Introduction

Multidisciplinary pain rehabilitation programs (PRPs) are effective in improving daily functioning

of patients with chronic low back pain (CLBP) 1-3. Most studies investigating the effects of PRPs

focused on the relationship between therapy content and effect. Guidelines for managing CLBP

are based on evidence obtained from these studies 4,5. However, this evidence might be biased.

In a literature search, we could not identify any study analyzing the relationship between dose

and effect as a primary objective. Two systematic reviews 2,3 were identifi ed that analyzed dose

and effect as a secondary objective. These reviews presented confl icting conclusions on dose

effects.

In a review of Guzman et al 2, 10 randomized controlled trials (RCTs) were included reporting

on 12 PRPs. PRPs were divided into 2 categories: daily intensive programs with more than 100 h

of therapy, and once- or twice-weekly programs with less than 30 h of therapy. Based the basis

of that distinction, it appeared that multidisciplinary PRPs of more than 100 h were superior to

monodisciplinary treatment, and PRPs of less than 30 h were not. The authors concluded that

intensive multidisciplinary PRPs are superior to less intensive multidisciplinary PRPs 2. These results

have been used in guidelines and clinical practice 4,6. However, it may be debated whether the

conclusion regarding dose is valid to support its clinical implications. First, there is an absence

of knowledge about PRPs with doses between 30 to 100 h. Second, no distinction was made

within intensive and less intensive PRPs, although there was a wide variety within the groups. Less

intensive PRPs ranged from 17.5 to 30 h, while intensive PRP ranged from 100 to 280 h. Third,

the review 2 was designed to assess the effect of different PRPs on clinically relevant outcomes.

All PRPs differed not only in dose, but also in content, setting and number of disciplines involved,

which may have confounded the dose aspect of the conclusion. Additionally, the authors

discussed whether the improvements gained with intensive PRP are worth the expense.

The review of Van Geen et al. 3 also had some limitations. It assessed the long-term effect of

PRPs with different contents and doses. Ten studies were included, and a distinction was made

between 30 h of training a week or more (intensive therapy) vs less than 30 h of training a week

(low-intensive therapy). Both the intensive and low-intensive PRPs showed positive effects. The

conclusion in that review was that the dose of the intervention had no substantial infl uence on

the effectiveness of the intervention.

In summary, while many studies provide evidence for effectiveness of multidisciplinary PRP, dose

of multidisciplinary PRP to achieve these effects remains unclear. The objective of the current

study was to analyze the infl uence of dose variables on the outcome of PRPs for patients with

CLBP.

R1R2R3R4R5R6R7R8R9


22 | Chapter 2

Methods

Publications were retrieved by computer-aided search on PUBMED, the Cochrane library, Cinahl,

and EMBASE up to October 2012. A specifi c search was developed using MeSH terms and other

relevant terms for each database. The PubMed search is described in the Supplement. Refworks

was used to store the results of the searches and to remove duplicates.

Selection of studies

Selection criteria were applied independently by two reviewers (FW and NH). The retrieved studies

were fi rst selected by title and abstract. Doubtful cases were discussed by the reviewers and

included or excluded for full text analysis by consensus. Full-text reports of studies eligible for

inclusion were analyzed. Disagreements were resolved by consensus or when necessary by a third

reviewer (HdV).

Studies were selected on the basis of the following inclusion criteria: (1) RCTs written in English,

Dutch or German; (2) the objective was to assess effectiveness of a multidisciplinary PRP for patients

with CLBP; PRP was defi ned as a rehabilitation program on the basis of the biopsychosocial model 7 with 3 or more disciplines providing the program (with or without a medical doctor); (3) total

number of contact hours of PRP was described; (4) participants were between 18 and 65 years

with disabling nonspecifi c CLBP for at least 3 months; (5) outcome variables were described in

the domain of disability, work participation or quality of life (QoL). Studies were excluded if: (1)

the multidisciplinary PRP was given in primary care; (2) the objective was to assess effectiveness

of only a biomedical intervention; (3) the study included participants who were diagnosed with

specifi c disorders or severe comorbidities interfering with PRP, such as heart failure, rheumatoid

arthritis, or psychiatric disorders; and (4) only total duration of PRP was reported.

Data management

Risk of bias of included studies was assessed according to the Cochrane Back Review Group 8

by 2 reviewers independently (FW and HdV). The criteria are presented in Table 1. Each criterion

was scored as positive (Y), negative (N) or unclear (U). The total score was computed by counting

the number of criteria scored as positive. Studies with a score of 6 or higher were defi ned as

low risk of bias; a score lower than 6 was defi ned as high risk of bias. In case of unclear scores,

corresponding authors of the studies were contacted by e-mail.

Performance bias was analysed with different items of blinding. The item regarding blinding of

care providers for intervention is frequently impossible in nondrug trials. To analyze the infl uence

of lack of blinding of care providers on the judgment of methodological quality, we performed a

sensitivity analyses by excluding item 4 of the risk of bias analyses.



2

Table 1 Criteria risk of bias analyses

Criteria risk of bias analyses1 Was the method of randomization adequate?2 Was the treatment allocation concealed?

Was knowledge of the allocated interventions adequately prevented during the study?3 Was the patient blinded to the intervention?4 Was the care provider blinded to the intervention?5 Was the outcome assessor blinded to the intervention?

Were incomplete outcome data adequately addressed?6 Was the drop-out rate described and acceptable?7 Were all randomized participants analyzed in the group to which they were allocated?8 Are reports of the study free of suggestion of selective outcome reporting?

Other sources of potential bias:9 Were the groups similar at baseline regarding the most important prognostic indicators?10 Were co-interventions avoided or similar?11 Was the compliance acceptable in all groups?12 Was the timing of the outcome assessment similar in all groups?

A data extraction form was developed and piloted before data extraction. Data were extracted

by 2 reviewers independently (FW and HdV). Disagreement was resolved by consensus or, if

necessary, by a third reviewer (MR). Data were extracted on general study, participant, and dose

characteristics (including total number of contact hours and total duration of the treatment

in weeks) and treatment content (including description of treatment, treatment components,

and number of disciplines). Reported outcome measures were categorized into disability, work

participation and QoL. The following interventions were distinguished: PRP, no treatment, care

as usual (CAU), individual physical treatment, individual psychological treatment, surgery, and

multidisciplinary treatment not defi ned as PRP.

For each intervention the effect size (ES) was calculated by subtracting the posttreatment mean

from the pretreatment mean (for each evaluation moment separately), divided by the pretreatment

standard deviation (SD). If the mean change (pretreatment - posttreatment) was reported, ES was

calculated by dividing the mean change by the SD of the change. If outcome was reported as a

proportion (eg, work participation), ES was calculated according to Hojat et al. 9. If medians and

ranges were reported 8,10-12 means and SDs were estimated 13. Because of the small sample sizes

of the different interventions, the ES were corrected with Hedges J as described by Borenstein et

al. 14.

To analyze the infl uence of dose variables on ES, a linear mixed effect model was applied

by SPSS version 18.0 software. Analyses were applied for all interventions on each outcome

category separately. The study of Hellum et al 15 was the only one comparing effects of PRP

with those of surgery. In that study, patients with CLBP were included, but it was not clear

R1R2R3R4R5R6R7R8R9


24 | Chapter 2

whether they had a specifi c or nonspecifi c diagnosis. Because of the unique character of that

study, the analyses were done twice: with and without this study for outcome categories disability

and QoL. ES per intervention was the response variable. Predictor variables were PRP (yes/no),

number of disciplines involved in the program, number of contact hours, treatment duration in

weeks, type of intervention (no treatment, CAU, surgery, individual physical therapy [exercise],

psychological treatment, other multidisciplinary treatment, PRP), evaluation moment (in months),

age, percentage of women and risk of bias (high/low). These variables were entered one by one

into the model. The variables remained in the model if the regression coeffi cient was signifi cant

(<0.05) or the model fi t increased signifi cantly (-2log-likelihood criterion). Random intercepts

were modeled because of a better fi t of the model. Patient characteristics such as duration of pain

and work status were reported insuffi ciently to be included in the statistical model.

Because augmentation of the content of the program can lead to deployment of more disciplines

and more contact hours or total treatment duration, we assumed that content and dose variables

could be correlated. Spearman correlations were used to analyze this assumption.

Results

Results of the search

A total of 2000 records was obtained from the electronic search, and 528 duplicates were

removed. In total, 1472 articles were assessed on title and abstract, and 127 articles were eligible

for full-text assessment. Grounds for exclusion after full-text assessment were mostly because

studies were not randomized, did not fi t the defi nition of multidisciplinary PRP, or contact hours

were not described. Follow-up studies of included RCTs were not regarded as a separate study,

but follow-up data were included in the analysis. Eighteen unique studies were included for risk

of bias assessment. (Figure 1)

Assessment of risk of bias

Agreement between reviewers expressed in Cohen’s Kappa was 0.65. All disagreements between

reviewers were solved by consensus. After fi rst assessment, 12 studies were categorized as

low risk of bias (score > 6) 16. However, many items were answered with “unclear” because

information was not clearly reported in the articles. Two articles had no unclear answers 17,18. The

e-mail address of 1 author 19 could not be retrieved. The authors of 14 articles were contacted for

clarifi cation, and answers were received from 11 of the 14 authors. After clarifi cation with the

author, 16 studies 8,10-12,15,17,18,20-28 had a score of 6 or higher (low risk of bias) and 2 studies 19,29

had a lower score and therefore a high risk of bias (Table 2).



2Articles retrieved by database search

n=2000

Total after duplicates removed

n=1472

Total eligible for full text evaluation

n=127

Studies included n=18

Duplicates n=528

Excluded on title/abstract selection

n=1345

Excluded on full text evaluation

n=109

Figure 1 Flow diagram of retrieval of studies

R1R2R3R4R5R6R7R8R9


26 | Chapter 2

In the sensitivity analysis, item 4 of the risk of bias analysis was omitted. Because no study scored

positive on this item, the total scores remained the same. Two studies had a score of 4 19,29. If the

cutoff point for low risk of bias would be decreased to 5 positive items, the risk of bias of these

2 studies would still be regarded as high.

Table 2 Results of risk of bias analyses

Author Year 1 2 3 4 5 6 7 8 9 10 11 12 Total + RoBAlaranta H 1994 ? ? ? ? ? + - + + ? ? + 4 highBendix AF 1995 + + - - + + - - + - ? + 6 lowBendix AF 1996 + + ? - + + - - + - ? + 6 lowBendix AF 1997 + + ? - + + - + + ? ? + 7 lowBendix T 2000 + + - - + + - + + ? ? + 7 lowJensen IB 2001 + + - - + + + + + ? + + 9 lowSkouen JS 2002 + + - - + + - + + ? + + 8 lowVollenbroek MMR 2004 + + - - + + + + + + + + 10 lowKaapa EH 2006 + + ? ? ? + + + + ? ? + 7 lowKainz B 2006 ? + ? ? ? - ? + + ? ? + 4 highSmeets RJ 2006 + + - - + + + + + + - + 9 lowLeeuw M 2008 + + + - - - + + + ? ? + 7 lowLinton SJ 2008 + + - - + - + + - ? ? + 6 lowMangels M 2009 + + ? ? - + + + + ? ? + 7 lowDufour N 2010 + + - - + + + + + + ? + 9 lowHellum C 2011 ? + - - + + + + - ? ? + 6 lowMeng K 2011 + + + - - + - + + ? ? + 7 lowRoche Leboucher G 2011 + + ? ? - + - + + + ? + 7 low

RoB= Risk of Bias, +=Yes, -=No, ?=Unclear

Description of studies

The included studies were published between 1995 and 2011. In total, 45 interventions

were compared (no treatment/CAU: 6; surgery: 1; individual physical therapy: 11; individual

psychological treatment: 1; other multidisciplinary treatment: 6; PRP: 20), and 27 comparisons

between interventions were made (Table 3). The total number of contact hours for the control

interventions (no treatment and CAU not included) ranged from 4.3 to 223.2 h (median=26 h,

IQR=15.0 - 46.0) and for PRPs from 6.4 to 196.8 h (median=120 h, IQR=51.8 - 135.0). Disability

was measured with 12 different measures (eg, RMDQ, PDI, subscales of SF-36). Six different

measures were included in category of work participation and consisted percentages of people

who were able to work (part time or full-time). QoL was measured with 4 different measures

(Euroqol, self-constructed questions regarding general well-being).



2

Tab

le 3

Ove

rvie

w o

f ch

arac

teris

tics

and

effe

ct s

izes

of

incl

uded

stu

dies

Aut

hor

Yr

Inte

rven

tions

nA

geSD

% W

om-

enD

urat

ion

pain

mth

s/yr

s(M

/Y)

SDTo

tal

dura

tion

Wks

/m

ths

Tota

l hr

sO

utco

me

dom

ain†

Eval

uatio

n m

omen

t (m

ths)

ES

Ala

rant

a H

1994

Cur

rent

Nat

iona

l Typ

e (C

NT)

of

reha

bilit

atio

n14

140

.44.

853

.2x

xx

3w

eeks

52.5

Dis

abili

tyk

pre

30.

498

pre

120.

487

AK

SELI

pro

gram

152

40.5

55.9

xx

x3

wee

ks12

6D

isab

ility

kpr

e3

0.90

0pr

e12

0.83

7Be

ndix

AF

1995

Inte

nsiv

e m

ultid

isci

plin

ary

trea

tmen

t pr

ogra

m +

func

tiona

l res

tora

tion

4040

*75

.0x

xx

3 +

3 F

Uw

eeks

135

Dis

abili

tya

pre

4x

Wor

kmpr

e4

1.07

2A

ctiv

e ph

ysic

al t

rain

ing

prog

ram

3144

*74

.2x

xx

6w

eeks

24D

isab

ility

apr

e4

xW

orkm

pre

40.

117

Act

ive

com

bine

d ps

ycho

-phy

sica

l pro

gram

3542

*77

.1x

xx

6w

eeks

24D

isab

ility

apr

e4

xW

orkm

pre

40.

362

Bend

ix A

F19

96In

tens

ive

mul

tidis

cipl

inar

y tr

eatm

ent

prog

ram

+ f

unct

iona

l res

tora

tion

4541

*71

.1x

xx

3 +

3 F

Uw

eeks

135

Dis

abili

tya

pre

4x

Wor

kmpr

e4

0.74

7N

o tr

eatm

ent

or t

reat

men

t el

sew

here

4940

*69

.4x

xx

0x

0D

isab

ility

apr

e4

xW

orkm

pre

40.

286

Bend

ix A

F19

97In

tens

ive

mul

tidis

cipl

inar

y tr

eatm

ent

prog

ram

+ f

unct

iona

l res

tora

tion

4640

*58

.7x

xx

3 +

3 F

Uw

eeks

135

Dis

abili

tya

pre

121.

087

Wor

kmpr

e12

1.41

3A

ctiv

e ph

ysic

al t

rain

ing

prog

ram

4343

*53

.5x

xx

6w

eeks

24D

isab

ility

apr

e12

0.17

0W

orkm

pre

120.

371

Act

ive

com

bine

d ps

ycho

-phy

sica

l pro

gram

4342

*58

.1x

xx

6w

eeks

24D

isab

ility

apr

e12

-0.3

30W

orkm

pre

120.

382

Bend

ix T

2000

Inte

nsiv

e m

ultid

isci

plin

ary

trea

tmen

t pr

ogra

m +

fun

ctio

nal r

esto

ratio

n59

40*

66.1

xx

x3

+ 3

FU

wee

ks13

8D

isab

ility

apr

e12

1.96

0

Wor

kmpr

e12

0.35

6O

utpa

tient

inte

nsiv

e ph

ysic

al t

rain

ing

prog

ram

6841

*64

.7x

xx

8w

eeks

36D

isab

ility

apr

e12

1.30

7

Wor

kmpr

e12

0.53

9Je

nsen

IB20

01Be

havi

oral

med

icin

e63

42.5

11.8

47.6

35.6

M68

.04

wee

ks13

4D

isab

ility

cpr

epo

st0.

514

60.

849

180.

388

Wor

kp18

1.53

5Be

havi

or o

rient

ed p

hysi

cal t

hera

py54

43.3

9.4

68.5

36.9

M67

.14

wee

ks80

Dis

abili

tyc

pre

post

0.34

96

0.92

618

0.28

1W

orkp

181.

365

R1R2R3R4R5R6R7R8R9


28 | Chapter 2

Aut

hor

Yr

Inte

rven

tions

nA

geSD

% W

om-

enD

urat

ion

pain

mth

s/yr

s(M

/Y)

SDTo

tal

dura

tion

Wks

/m

ths

Tota

l hr

sO

utco

me

dom

ain†

Eval

uatio

n m

omen

t (m

ths)

ES

Cog

nitiv

e be

havi

oral

the

rapy

4943

.89.

644

.922

.7M

38.4

4w

eeks

54D

isab

ility

cpr

epo

st0.

136

60.

379

180.

112

Wor

kp18

1.41

1C

ontr

ol g

roup

/tre

atm

ent

as u

sual

4843

.910

.858

.327

.3M

39.2

0x

0D

isab

ility

cpr

epo

st0.

037

60.

406

180.

166

Wor

kp18

1.29

8Sk

ouen

JS

2002

Exte

nsiv

e m

ultid

isci

plin

ary

trea

tmen

t57

42.9

10.5

70.2

xx

x4

wee

ks12

0W

orkn

120.

802

180.

522

240.

558

Ligh

t m

ultid

isci

plin

ary

trea

tmen

t52

43.7

11.5

59.6

xx

xx

x4.

25W

orkn

120.

807

180.

511

240.

481

Con

trol

s86

4411

.764

.0x

xx

0x

0W

orkn

120.

326

180.

385

240.

462

Volle

nbro

ek M

MR

2004

Roes

sing

h Ba

ck R

ehab

ilita

tion

Prog

ram

7938

.59.

872

*M

7w

eeks

63D

isab

ility

dpr

epo

st0.

495

40.

743

Qol

spr

epo

st0.

147

40.

367

Usu

al C

are

8439

.59.

948

*M

xx

xD

isab

ility

dpr

epo

st0.

000

40.

495

Qol

spr

epo

st-0

.114

40.

228

Kaa

pa E

H20

06M

ultid

isci

plin

ary

reha

bilit

atio

n64

467.

910

0x

xx

8w

eeks

70D

isab

ility

epr

epo

st0.

420

60.

467

120.

607

240.

527

Wor

kqpr

epo

st0.

609

60.

419

120.

419

240.

377

QoL

tpo

stx



2

Aut

hor

Yr

Inte

rven

tions

nA

geSD

% W

om-

enD

urat

ion

pain

mth

s/yr

s(M

/Y)

SDTo

tal

dura

tion

Wks

/m

ths

Tota

l hr

sO

utco

me

dom

ain†

Eval

uatio

n m

omen

t (m

ths)

ES

Indi

vidu

al P

hysi

othe

rapy

6646

.57.

010

0x

xx

6 to

8w

eeks

10D

isab

ility

epr

epo

st0.

186

60.

491

120.

448

240.

377

Wor

kqpr

epo

st0.

283

60.

354

120.

318

240.

035

QoL

tpo

stx

Kai

nz B

2006

Med

ical

tra

inin

g th

erap

y (M

TT)

100

47.9

x50

xx

x15

wee

ks27

.4D

isab

ility

lpr

e6

0.34

7W

orkr

pre

60.

417

Qol

vpr

epo

st0.

574

60.

485

Out

patie

nt P

hysi

othe

rapy

(EA

P)10

047

.9x

59.9

xx

x8

wee

ks46

Dis

abili

tyl

pre

60.

255

Wor

krpr

e6

0.54

3Q

olv

pre

post

0.56

06

0.48

0O

utpa

tient

Reh

abili

tatio

n (A

R)10

047

.9x

50x

xx

5w

eeks

76.4

Dis

abili

tyl

pre

60.

335

Wor

krpr

e6

0.29

3Q

olv

pre

post

0.61

06

0.46

0Sm

eets

RJ

2006

Act

ive

phys

ical

tre

atm

ent

5342

.79.

141

0556

.9M

75.9

10W

eeks

43.5

Dis

abili

tyd

pre

post

0.71

46

0.33

012

0.33

4C

ogni

tive

beha

vior

al t

reat

men

t58

42.5

9.7

58.6

68.3

M74

.210

Wee

ks26

.5D

isab

ility

dpr

epo

st0.

791

60.

357

120.

360

Com

bine

d tr

eatm

ent

6140

.710

.137

.756

.1M

67.5

10W

eeks

69.5

Dis

abili

tyd

pre

post

0.53

36

0.35

412

0.34

1W

aitin

g lis

t51

40.6

11.2

5143

.8M

70.8

10W

eeks

0D

isab

ility

dpr

epo

st0.

020

6x

12x

R1R2R3R4R5R6R7R8R9


30 | Chapter 2

Aut

hor

Yr

Inte

rven

tions

nA

geSD

% W

om-

enD

urat

ion

pain

mth

s/yr

s(M

/Y)

SDTo

tal

dura

tion

Wks

/m

ths

Tota

l hr

sO

utco

me

dom

ain†

Eval

uatio

n m

omen

t (m

ths)

ES

Leeu

w20

08Ex

posu

re in

viv

o tr

eatm

ent

4246

.45

9.33

52.4

8.13

Y9.

95x

x16

Dis

abili

tyh,

ipr

epo

st1.

492

61.

231

Ope

rant

gra

ded

activ

ity43

44.2

19.

5444

.29.

00Y

8.94

xx

26D

isab

ility

h,i

pre

post

0.73

86

0.71

9Li

nton

SJ

2008

Expo

se T

AU

(tre

atm

ent

as u

sual

)21

469.

962

83%

>Y

xx

x13

Dis

abili

tyi

pre

post

0.47

3W

aitin

g lis

t TA

U (t

reat

men

t as

usu

al)

2549

7.3

4876

% >

Yx

xx

0D

isab

ility

ipr

epo

st0.

006

Man

gels

M20

09Tr

aditi

onal

ort

hope

dic

reha

bilit

atio

n tr

eatm

ent

131

48.7

14.7

78.6

xx

x3

wee

ks19

6.8

Dis

abili

tyg

pre

post

0.30

1

120.

333

QoL

upr

epo

st0.

419

120.

268

Mul

tidis

cipl

inar

y re

habi

litat

ion

113

49.5

9.0

78.8

xx

x4

wee

ks22

3.2

Dis

abili

tyg

pre

post

0.48

712

0.34

4Q

oLu

pre

post

0.55

012

0.39

6M

ultid

isci

plin

ary

reha

bilit

atio

n +

bo

oste

r se

ssio

ns11

948

.315

.875

.6x

xx

4 +

12 FU

wee

ks +

FU

mth

s22

3.2

Dis

abili

tyg

pre

post

0.37

012

0.30

6Q

oLu

pre

post

0.49

012

0.51

1D

ufou

r N

2010

Gro

up b

ased

mul

tidis

cipl

inar

y bi

opsy

chos

ocia

l reh

abili

tatio

n pr

ogra

m12

946

.653

.456

.6x

xx

12w

eeks

87D

isab

ility

b,r

pre

post

0.60

26

0.51

512

0.45

524

0.54

9W

orkm

pre

post

0.23

124

0.39

1In

tens

ive

indi

vidu

ally

the

rapi

st-a

ssis

ted

back

m

uscl

e st

reng

then

ing

exer

cise

s14

349

.750

.355

.9x

xx

12w

eeks

46D

isab

ility

b,r

pre

post

0.35

96

0.28

912

0.21

124

0.23

6W

orkm

pre

post

0.04

124

0.36

2H

ellu

m C

2011

Reha

bilit

atio

n87

40.8

7.1

58.6

xx

x3

to 5

wee

ks60

Dis

abili

tyb,

epr

e1.

51.

341

31.

144

60.

726

121.

379



2

Aut

hor

Yr

Inte

rven

tions

nA

geSD

% W

om-

enD

urat

ion

pain

mth

s/yr

s(M

/Y)

SDTo

tal

dura

tion

Wks

/m

ths

Tota

l hr

sO

utco

me

dom

ain†

Eval

uatio

n m

omen

t (m

ths)

ES

241.

527

QoL

s1.

50.

894

30.

671

60.

766

120.

862

241.

054

Surg

ery

8641

.17.

146

.5x

xx

xx

xD

isab

ility

b,e

pre

1.5

1.12

13

1.78

76

1.87

012

2.07

624

2.13

1Q

oLs

1.5

0.95

73

1.32

06

1.25

412

1.22

124

1.25

4M

eng

K20

11N

ew b

iops

ycho

soci

al b

ack

scho

ol p

rogr

am19

750

.27.

665

.251

.5%

>5

Yx

24.3

(4

.8)

mea

n da

ys6.

4D

isab

ility

jpr

e6

0.70

2

120.

377

Trad

ition

al b

ack

scho

ol p

rogr

am (C

AU

)18

549

.57.

763

39.1

%>

5 Y

x24

.9

(5.2

)m

ean

days

5.5

Dis

abili

tyj

pre

60.

569

120.

420

Roch

e Le

bouc

he

G20

11Fu

nctio

nal r

esto

ratio

n pr

ogra

m68

40.8

7.4

32.4

xx

x5

wee

ks15

0D

isab

ility

fpr

e12

1.74

7W

orko

pre

121.

037

Act

ive

indi

vidu

al t

hera

py64

38.7

6.1

37.5

xx

x5

wee

ks15

Dis

abili

tyf

pre

120.

652

Wor

kopr

e12

0.75

1

*=m

edia

n, x

=no

dat

a av

aila

ble.

ES: E

ffec

t si

ze. F

U: F

ollo

w u

p, Q

oL: Q

ualit

y of

Life

.†

Mea

sure

men

t in

stru

men

t us

ed f

or o

pera

tiona

lizat

ion

of o

utco

me

dom

ain.

a:

Func

tion.

b:

SF36

phy

sica

l com

pone

nt.

c: S

F36

phys

ical

fun

ctio

ning

+ S

F36

role

ph

ysic

al.

d: R

MD

Q.

e: O

swes

try.

f:

Pain

impa

ct.

g: P

DI.

h: P

CS.

i: Q

BPD

S. j:

Phy

sica

l act

ivity

(fr

eibu

rger

). k:

Mill

ion

Inde

x. l:

Fun

ktio

nsfa

higk

eit.

m:

wor

k re

adin

ess.

n:

Ret

urn

to w

ork.

o:

perc

enta

ge f

ull t

ime

at w

ork.

p:

wor

k pa

rtic

ipat

ion.

q:

subj

ectiv

e w

ork

capa

city

. r:

arbe

idsf

ähig

keit.

s:

Euro

qol.

t: g

ener

al w

ellb

eing

. u:

life

sa

tisfa

ctio

n qu

estio

nnai

re. v

: som

atis

che

gesu

ndhe

it.

R1R2R3R4R5R6R7R8R9


32 | Chapter 2

ES for disability ranged from -0.33 to 2.13 (median 0.36, IQR 0.23 - 0.67) for control interventions

and from 0.31 to 1.96 (median 0.54, IQR 0.43 - 1.04) for PRPs. ES for work participation ranged

from 0.04 to 1.41 (median 0.38, IQR 0.32 - 0.54) for controls and for PRPs from 0.23 to 1.54

(median 0.54, IQR 0.38 - 0.98). QoL ranged from -0.11 to 1.32 (median 0.56, IQR 0.34 - 1.24)

for controls and from 0.15 to 1.05 (median 0.55, IQR 0.43 - 0.81) for PRPs (Table 4). In one study 15, the control intervention (surgery) was superior to PRP. That study was also responsible for the

highest ES for the control intervention for disability (2.13) and QoL (1.25).

Table 4 Ranges of ES per outcome category for PRPs and control interventions

Outcome PRP Controls PRP ControlsIncluding study of Hellum et al. Excluding study of Hellum et al.

Disability 0.306 to 1.960 -0.330 to 2.131 0.306 to 1.960 -0.330 to 1.307Work Participation 0.231 to 1.535 0.035 to 1.411 0.231 to 1.535 0.035 to 1.411Quality of life 0.147 to 1.054 -0.114 to 1.320 0.147 to 0.610 -0.114 to 0.574

ES: Effect size. PRP: Pain Rehabilitation Program.

Mixed effect modeling

Disability

Type of intervention, evaluation moment, and percentage of women signifi cantly contributed to

the regression equation. In this analysis, the reference category was effect of PRP immediately

after treatment, with 0% women as participants. The negative coeffi cients indicate that control

interventions had a smaller ES than PRP (β=0.501), except surgery, which had a signifi cantly

larger ES (β=0.748), indicating that the ES of surgery was 0.501 + 0.748= 1.249. The coeffi cient

of 12 to 24 months for evaluation moment indicates that the ES is 0.407 points larger at 12 to

24 months than immediately after intervention (reference category) (Table 5). The coeffi cient

of 0.002 for percentage of women indicates that for every 1% women extra as participants,

the ES increases with 0.002. Performing the analyses without the study of Hellum et al 15, the

reference category was effect of PRP with 6 weeks’ duration after treatment, with 0% women as

participants. A signifi cant contribution of type of intervention, evaluation moment, percentage

of women, and duration in weeks was found (Table 5). All control interventions had a smaller

ES than PRP (β=0.429) ranging from -0.714 to -0.154. Evaluation moment between 12 and 24

months after the study had a regression coeffi cient of 0.447, indicating that the ES at 12 to 24

months is 0.447 larger than immediately after treatment. The coeffi cient of 0.003 for percentage

of women indicates that for every 1% women extra as participants, the ES increases by 0.003.

The coeffi cient of -0.037 for duration in weeks indicates that for each additional week above 6

weeks, the ES decreases with 0.037.



2

Table 5 Results of mixed effect model analyses for the different outcome variables, in and excluding data from the study of Hellum et al (2011)

Outcome Variable β SE β pDisability a Intercept r 0.501 0.167 0.004

No treatment -0.617 0.383 0.109CAU -0.719 0.176 <0.001Surgery 0.748 0.218 0.001Individual therapy/exercise -0.142 0.108 0.188Psychological treatment -0.496 0.176 0.006Other multidisciplinary treatment -0.294 0.198 0.1401.5 – 6 months FU 0.173 0.116 0.13912 months FU 0.182 0.142 0.20112-24 months FU 0.407 0.131 0.002Percentage women 0.002 0.001 0.086

Work Participation b Intercept r 0.475 0.116 0.001Number of disciplines 0.088 0.025 0.001

Quality of life c PRP r 0.538 0.104 0.006CAU -0.241 0.112 0.041Surgery 0.372 0.071 <0.001Individual therapy/exercise -0.011 0.096 0.912Other multidisciplinary treatment -0.147 0.099 0.149

Analyses excluding study of Hellum et al.(2011)Disability d Intercept r 0.429 0.160 0.009

Individual therapy/exercise -0.154 0.120 0.206Psychological treatment -0.414 0.189 0.031Other multidisciplinary treatment -0.714 0.332 0.0341.5 – 6 months FU 0.174 0.152 0.25512 months FU 0.188 0.158 0.23712-24 months FU 0.447 0.171 0.010Percentage women 0.003 0.002 0.108Duration in weeks -0.037 0.018 0.038

Quality of life e Intercept r 0.417 0.033 <0.001CAU -0.231 0.087 0.018Individual therapy/exercise 0.040 0.058 0.496Other multidisciplinary treatment -0.160 0.075 0.050Age (centered at 45 years) 0.023 0.007 0.006

a: Reference category: effect of PRP post treatment, with 0% women as participants. b: Reference category: effect of treatment with 0 disciplines. c: Reference category: PRP. d: Reference category: effect of PRP with 6 weeks duration, post treatment, with 0% women as participants. e: Reference category: effect of PRP for patients at age of 45 years. β: Regression coeffi cient. SE: standard error β. r: random. CAU: Care as usual. PRP: Pain rehabilitation Program. FU: follow up/evaluation moment. *Although β not signifi cant model fi t increased signifi cantly.

Work Participation

The number of disciplines signifi cantly contributed to the regression equation. All other variables

entered into the model did not contribute signifi cantly. Table 5 shows that every discipline added

R1R2R3R4R5R6R7R8R9


34 | Chapter 2

to the intervention, compared to the reference category of treatment with 0 disciplines, will

increase the ES with 0.088.

QoL

To predict the ES for outcome QoL, only type of intervention contributed signifi cantly to the

regression equation. Table 5 shows an effect of 0.417 for PRP (reference category). The negative

coeffi cients indicate a smaller effect for control interventions except for the intervention surgery

(β=0.372), indicating an effect for surgery of 0.538 + 0.372= 0.910. Type of intervention and

age contributed signifi cantly to the regression equation when performing the analysis without

the data of Hellum et al. 15 (Table 5). For QoL, the effect of PRP for patients at age of 45 years is

0.417; CAU and other multidisciplinary treatment for patients at age 45 years had smaller effect;

only individual therapy for patients aged 45 years showed a 0.040 (not signifi cant) larger effect

compared to PRP. Contact hours and number of disciplines were correlated (r=0.57, p<.01). The

number of disciplines involved and contact hours were larger in PRP (respectively, median 3 and

87.0) compared to other interventions (respectively, median 1 and 27.4). No other signifi cant

associations were found between contact hours, total duration of the treatment and number of

disciplines involved.

Discussion

To our knowledge, this is the fi rst systematic review analyzing the relationship between dose and

effect of multidisciplinary PRPs. Twelve out of the 18 studies were unique; they were not included

in the reviews of Guzman et al. 2 and Van Geen et al. 3. Some studies were published after these

2 reviews; other studies were not included in the current review because of different inclusion

and exclusion criteria. The current review shows a wide variety in contact hours and ES for all

interventions treating patients with CLBP. Within the interventions defi ned as multidisciplinary

PRP, substantial variation was observed in ES, contact hours, content, total duration and number

of disciplines. Analyzing the infl uence of predictor variables on effect of PRP on disability

demonstrated that the type of intervention, evaluation moment, percentage of women, and

the total duration in weeks infl uenced that effect. The number of disciplines involved infl uenced

effects on work participation, and QoL was additionally infl uenced by the type of intervention

and age of participants. PRP and surgery had the strongest effects on the outcome disability. PRP

had the strongest effect on the outcome work participation. Surgery, as described by Hellum et al. 15, was the only intervention that was superior to PRP. Surprisingly, no other studies were obtained

from the literature search comparing surgery to PRP for patients with CLBP.



2

The review of Guzman et al. 2 provided evidence that intensive PRP of more than 100 h of therapy

produced greater improvements in function than PRP of less than 30 h of therapy. The current

review provides a wide variety of contact hours in PRPs, with a range of 6.4 to 196.8 h, with

larger ES than most of the control interventions. This shows that there are also PRPs producing

improvements within the range of 30 to 100 h of therapy. Therapies within this range of contact

hours were not included in the review of Guzman et al. 15,17,18,22,26. It is unknown whether these

PRPs are as effective as the PRPs of more than 100 h.

This study focused on dose aspects of treatment. The results of the analyses of this review showed

that type of intervention, evaluation moment, percentage of women, number of disciplines,

duration in weeks, and the age of the participants infl uenced the effect of interventions for

patients with CLBP. However, because contact hours and number of disciplines were interrelated

for all interventions, we are not able to draw conclusions about the relationship between

independent components and effects. In most cases, the number of disciplines involved depends

on the aim and content of the treatment. However, we demonstrated that it is currently not

possible to disentangle dose and content, which supports the importance of paying attention

to both components for analyzing the effectiveness of PRPs. Regarding the evaluation moment,

it is not clear whether the effect increases over time as a result of the intervention or as a result

of natural recovery. PRPs were on average more effective than the other interventions, but all

included studies analyzed PRPs with different doses and contents. Because studies were not

designed to investigate dose variables independently, we were not able to analyze the relationship

between dose and effects in PRP alone. Consequently, research on the effectiveness of PRPs may

have been biased because it has been focused solely on content and neglected dose of treatment.

All included studies were similar with regard to participants and study design. Subgroups were

made to analyze outcome domains and follow-up time separately. Therefore, we judged the

studies to be suffi ciently homogeneous to perform a meta-analysis. The heterogeneity of the

content of PRPs and outcomes is a strength of the study when focusing on the clinical use of the

information about dose aspects of PRPs, but is also a weakness regarding the methodological

strength of this review.

Study limitations

RCTs included in this study were aimed to analyze the effect of PRP. No RCTs were obtained

analyzing the relationship between dose and effect. Therefore, this study provides only indirect

conclusions, which is a limitation of the study. Another limitation is the statistical heterogeneity,

which was almost inevitable because of the wide variety of PRPs in terms of content, outcome

measures, dose and so on. Fortunately, more and more guidelines and recommendations 30 are

being published to encourage the homogeneity in chronic pain research. Another shortcoming

R1R2R3R4R5R6R7R8R9


36 | Chapter 2

of this study is that the conclusions are based on a small sample of RCTs concerning PRPs,

resulting in a limited power. This small sample was the result of selection criteria for studies

to provide similar interventions and study designs and from the lack of consistent reporting of

dose variables. The small sample resulted in categories of evaluation moment with different

intervals and ES from different interventions, which limits the interpretability of these results.

The coeffi cients of the categories seem to indicate a much larger effect at 12 to 24 months (β=

0.460, β= 0.485) compared to the other evaluation moments. However, this category contains 12

months compared to 1 or 4.5 months interval in the other categories. An increase of ES during

the different evaluation moments exists, but it fl attens. Risk of bias was assessed with the tool

recommended by the Cochrane Back Review Group. The choice to use this risk of bias scale may

be debated because of the ongoing discussion about the blinding of patients and therapists

in nondrug trials. Although blinding is recommended to assess performance bias in RCTs 31,

patients were blinded in only 2 studies, and care providers were blinded in none of the studies 23,27. Low risk of bias was defi ned as scoring positively for 6 or more out of 12 items. Blinding of

patients and care providers covers 2 out of 12 items. Consequently, failure to blind care providers

and/or patients does not necessarily lead to a low quality rating. To fulfi ll the criteria of risk

of bias, there are methods regarding blinding of patients to interventions in nondrug trials,

such as evaluating treatment credibility, patient expectations and patient treatment preference.

Blinding of care providers is more diffi cult in nondrug trials; nevertheless, knowledge of which

intervention a patient received can infl uence treatment effects. Because of this risk of bias, the

extension of the CONSORT statement for nondrug trials suggests to report the interpretation of

the results regarding lack of or partial blinding in RCTs 32. Another option to control for lack of

blinding is measuring care providers’ allegiance and expectations towards dose and content of

the treatment protocol. To analyze the effect of the risk of bias analysis on the outcome of this

review, we performed a sensitivity analysis. Taking the item blinding of care providers out of the

risk of bias analysis would not have made a difference in the overall judgment of high or low risk

of bias, or of the fi nal conclusions of this review. Other quality assessment tools to analyze the risk

of bias are available for trials in which it is diffi cult to blind participants or therapists 33,34. Another

limitation is the language restriction. It is possible that other high-quality RCTs exist written in

other languages that could provide relevant information regarding aspects of dose.

Conclusion

This review supports the evidence that PRPs are effective and superior to most other interventions

in treating patients with CLBP on disability and work participation. Within PRPs, a wide variety of

contents and doses are observed. Dose aspects of PRPs are reported and analyzed insuffi ciently. On

the basis of the current literature, it is unknown how many hours, months, or weeks are needed

to achieve the best effects. It is also not possible to draw conclusions about the infl uences of dose

on effect, which could support the clinical decisions for dose of PRP for individual patients. To



2

further improve PRPs, there is need to report dose aspects (total duration of treatment, number of

contact hours, number of disciplines) in analyzing effects of PRPs and to study dosage (duration

and intensity) of PRPs. Additionally, qualitative studies aiming to identify relevant aspects of

dosage of PRPs could provide more insight in the importance and clinical relevance of dose.

Acknowledgement

We certify that no party having a direct interest in the results of the research supporting this

article has or will confer a benefi t on us or on any organization with which we are associated.

R1R2R3R4R5R6R7R8R9


38 | Chapter 2

References




4. Airaksinen O, Brox JI, Cedraschi C, et al. Chapter 4. european guidelines for the management of chronic nonspecifi c low back pain. Eur Spine J. 2006;15 Suppl 2:S192-300.

5. Sanders SH, Harden RN, Vicente PJ. Evidence-based clinical practice guidelines for interdisciplinary rehabilitation of chronic nonmalignant pain syndrome patients. Pain Pract. 2005;5(4):303-315.

6. van Hooff ML, van der Merwe JD, O’Dowd J, et al. Daily functioning and self-management in patients with chronic low back pain after an intensive cognitive behavioral programme for pain management. Eur Spine J. 2010;19(9):1517-1526.

7. Waddell G. 1987 volvo award in clinical sciences. A new clinical model for the treatment of low-back pain. Spine (Phila Pa 1976). 1987;12(7):632-644.

8. Bendix AF, Bendix T, Vaegter K, Lund C, Frolund L, Holm L. Multidisciplinary intensive treatment for chronic low back pain: A randomized, prospective study. Cleve Clin J Med. 1996;63(1):62-69.

9. Hojat M, Xu G. A visitor’s guide to effect sizes: Statistical signifi cance versus practical (clinical) importance of research fi ndings. Adv Health Sci Educ Theory Pract. 2004;9(3):241-249.

10. Bendix AF, Bendix T, Lund C, Kirkbak S, Ostenfeld S. Comparison of three intensive programs for chronic low back pain patients: A prospective, randomized, observer-blinded study with one-year follow-up. Scand J Rehabil Med. 1997;29(2):81-89.

11. Bendix AF, Bendix T, Ostenfeld S, Bush E, Andersen. Active treatment programs for patients with chronic low back pain: A prospective, randomized, observer-blinded study. Eur Spine J. 1995;4(3):148-152.

12. Bendix T, Bendix A, Labriola M, Haestrup C, Ebbehoj N. Functional restoration versus outpatient physical training in chronic low back pain: A randomized comparative study. Spine (Phila Pa 1976). 2000;25(19):2494-2500.

13. Hozo SP, Djulbego. Estimating the mean and variance from the median, range, and the size of a sample. BMC Medical research methodology. 2005;5:13.

14. Borenstein M. Introduction to meta-analysis. Chichester, UK. Wiley; 2009:421.

15. Hellum C, Johnsen LG, Storheim K, et al. Surgery with disc prosthesis versus rehabilitation in patients with low back pain and degenerative disc: Two year follow-up of randomised study. BMJ. 2011;342:d2786.

16. Furlan AD, Pennick V, Bombardier C, van Tulder M, Editorial Board, Cochrane Back Review Group. 2009 updated method guidelines for systematic reviews in the cochrane back review group. Spine (Phila Pa 1976). 2009;34(18):1929-1941.




2

18. Vollenbroek-Hutten MM, Hermens HJ, Wever D, Gorter M, Rinket J, Ijzerman MJ. Differences in outcome of a multidisciplinary treatment between subgroups of chronic low back pain patients defi ned using two multiaxial assessment instruments: The multidimensional pain inventory and lumbar dynamometry. Clin Rehabil. 2004;18(5):566-579.

19. Alaranta H, Rytokoski U, Rissanen A, et al. Intensive physical and psychosocial training program for patients with chronic low back pain. A controlled clinical trial. Spine (Phila Pa 1976). 1994;19(12):1339-1349.

20. Jensen IB, Bergstrom G, Ljungquist T, Bodin L, Nygren AL. A randomized controlled component analysis of a behavioral medicine rehabilitation program for chronic spinal pain: Are the effects dependent on gender? Pain. 2001;91(1-2):65-78.

21. Skouen JS, Grasdal AL, Haldorsen EM, Ursin H. Relative cost-effectiveness of extensive and light multidisciplinary treatment programs versus treatment as usual for patients with chronic low back pain on long-term sick leave: Randomized controlled study. Spine (Phila Pa 1976). 2002;27(9):901-9; discussion 909-10.

22. Kaapa EH, Frantsi K, Sarna S, Malmivaara A. Multidisciplinary group rehabilitation versus individual physiotherapy for chronic nonspecifi c low back pain: A randomized trial. Spine (Phila Pa 1976). 2006;31(4):371-376.

23. Leeuw M, Goossens ME, van Breukelen GJ, et al. Exposure in vivo versus operant graded activity in chronic low back pain patients: Results of a randomized controlled trial. Pain. 2008;138(1):192-207.

24. Linton SJ, Boersma K, Jansson M, Overmeer T, Lindblom K, Vlaeyen JW. A randomized controlled trial of exposure in vivo for patients with spinal pain reporting fear of work-related activities. Eur J Pain. 2008;12(6):722-730.

25. Mangels M, Schwarz S, Worringen U, Holme M, Rief W. Evaluation of a behavioral-medical inpatient rehabilitation treatment including booster sessions: A randomized controlled study. Clin J Pain. 2009;25(5):356-364.

26. Dufour N, Thamsborg G, Oefeldt A, Lundsgaard C, Stender S. Treatment of chronic low back pain: A randomized, clinical trial comparing group-based multidisciplinary biopsychosocial rehabilitation and intensive individual therapist-assisted back muscle strengthening exercises. Spine (Phila Pa 1976). 2010;35(5):469-476.

27. Meng K, Seekatz B, Roband H, Worringen U, Vogel H, Faller H. Intermediate and long-term effects of a standardized back school for inpatient orthopedic rehabilitation on illness knowledge and self-management behaviors: A randomized controlled trial. Clin J Pain. 2011;27(3):248-257.

28. Roche-Leboucher G, Petit-Lemanac’h A, Bontoux L, et al. Multidisciplinary intensive functional restoration versus outpatient active physiotherapy in chronic low back pain: A randomized controlled trial. Spine (Phila Pa 1976). 2011;36(26):2235-2242.

29. Kainz B, Gulich M, Engel EM, Jackel WH. Comparison of three outpatient therapy forms for treatment of chronic low back pain-- fi ndings of a multicentre, cluster randomized study. Rehabilitation (Stuttg). 2006;45(2):65-77.

30. Dworkin RH, Turk DC, Farrar JT, et al. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain. 2005;113(1-2):9-19.

31. Higgins JP, Altman DG, Gotzsche PC, et al. The cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928.

32. Boutron I, Moher D, Altman DG, Schulz KF, Ravaud P, CONSORT Group. Extending the CONSORT statement to randomized trials of nonpharmacologic treatment: Explanation and elaboration. Ann Intern Med. 2008;148(4):295-309.

R1R2R3R4R5R6R7R8R9


40 | Chapter 2

33. Yates SL, Morley S, Eccleston C, de C Williams AC. A scale for rating the quality of psychological trials for pain. Pain. 2005;117(3):314-325.

34. Boutron I, Moher D, Tugwell P, et al. A checklist to evaluate a report of a nonpharmacological trial (CLEAR NPT) was developed using consensus. J Clin Epidemiol. 2005;58(12):1233-1240.



2

FrFrFranaankakaka PPPC CC WaWaWatetetersrsr chchchoooot,t,t PPieieeteteter r UUU DiDiD jkjkksttstraraa PPPhDhDhD,JaJaJannn HBHBHB GGGeeeeeertrtrtzezez nnn MDMDMD,, PhPhPhD,D,D, MMMicicichihihielele FFF RRenene ememanann PPPhDhDhD.

InInInt tt J JJ ReReRehahahabiibil ll ReReRes ss 20202015515, 383838 ((1)1)1): 343434-9-9-9.

R1R2R3R4R5R6R7R8R9


44 | Chapter 3

Abstract

The aim of this study was to analyze the course of reduction of disability during a pain rehabilitation

program (PRP) and factors infl uencing this course. A prospective cohort study was carried out. All

patients with chronic musculoskeletal pain treated in a PRP between March 2010 and December

2010 were eligible for this study. All patients were treated at a University-based rehabilitation

center and received an outpatient multidisciplinary PRP. Main outcome measures, Pain Disability

Index (PDI), and average pain measured with a numeric rating scale, were measured every 2

weeks during the PRP. To analyze the course of disability, a linear mixed-effect model was applied.

One hundred twenty-eight patients participated in the study, of which 20% dropped out during

the PRP. Initial PDI (β= 0.8), treatment week (β= -0.2), treatment week squared (β=0.03), average

pain (β= 2.3), and interaction between initial PDI and treatment week (β= -0.02) infl uenced the

course of disability during PRP. Disability reduces during the PRP. Initial PDI, treatment week,

average pain, and interaction between initial PDI and treatment week infl uence the course of

disability reduction during the PRP. These results could aid in predicting the required duration of

a PRP at the start.

Keywords

Chronic pain, disability, Pain Disability Index, Patient Care Team, Rehabilitation


Course of disability reduction during a PRP | 45

3

Introduction

On the basis of pre and post measurements, it is known that pain rehabilitation programs (PRPs)

are effective in disability reduction in patients with chronic musculoskeletal pain (CMP) 1-7.

However, the course of disability reduction is unknown. Understanding the course of disability

reduction could aid in determining the optimum duration of PRPs. This understanding would

benefi t patients, clinicians, insurance companies and health care providers because it prevents

overtreatment or undertreatment and contributes towards better use of (public) recourses.

A wide diversity of content, composition, and duration of PRPs exist 3,4. In one systematic review,

differences in duration were unrelated to differences in outcomes 3. However, in another review

it was reported that multidisciplinary PRP of more than 100 h was superior to monodisciplinary

treatment, and that multidisciplinary PRP of less than 30 h was not superior to monodisciplinary

treatment. Although the authors focused on the content of different PRPs and the review was

designed to assess the effect of PRP, they concluded that PRPs of more than 100 h were more

effective. The conclusions of these reviews were used to establish guidelines and clinical practice 8,9. However, these conclusions were based only on pre-post assessments. To our knowledge,

no studies have analyzed the course of disability reduction during the programs to establish the

optimum duration of a PRP. In addition, it is unknown whether the improvements gained with

intensive PRPs are worth the expenses 4. Theoretically, if 90% of the improvements are observed

in the fi rst half of the PRP, it may be debated whether these 100 h or more, as proposed by

Guzman et al 4, are needed.

Because no previous research is available into this specifi c aspect of PRPs, we hypothesized that

there would be several possible courses of reduction of disability (Fig. 1). Line A assumes that the

course of disability has a steeper slope at the end of the PRP. This implies that disability decreases

most in the last part of the PRP. In contrast, line B assumes a steeper slope at the fi rst half of

the PRP. This Implies that disability decreases most in the fi rst part of the PRP. Line C implies a

linear decrease of disability, and an alternating pattern of decrease and increase during the PRP

is hypothesized in line D.

The aim of this study was to analyze the course of reduction of disability during PRP and to

analyze factors infl uencing this course.

R1R2R3R4R5R6R7R8R9


46 | Chapter 3

Treatment weeks

Mean PDI

1 2 3 X

A

B

C

D

Figure 1 Hypothesized courses of reduction of disability during a hypothetical Pain Rehabilitation ProgramLine A: the decrease of disability has a steeper slope at the second half of PRP. Line B: the decrease of disability is steeper slope at the fi rst half of PRP. Line C: a linear decrease of disability during PRP.Line D: an alternating pattern of decrease and increase of disability during PRP.

Patients and methods

Study design and setting

A prospective cohort study was carried out with patients with CMP admitted to the outpatient

multidisciplinary PRP of the Center for Rehabilitation of the University Medical Center Groningen

(UMCG).

The primary aim of this PRP, on the basis of cognitive behavioral principles, was to decrease

pain-related disability. The rehabilitation team consists of physicians, occupational therapists,

physiotherapists and psychologists. PRP consists of education on differences between nociceptive

pain and chronic pain. Patients are counseled to refl ect on their own pain management strategies

(avoiding pain) and how these strategies could be changed into more healthy management

strategies (coping with pain, alternation between physical activity and rest, and to gradually

increase functioning). Each patient sets individual treatment goals to create a meaningful life



3

despite the pain. To achieve these goals, each professional applies specifi c techniques. Physical

therapists apply exercise programs and sports activities, improve patients’ confi dence in movement,

and reduce pain related fear. Occupational therapists assess current activities and patterns in daily

living and educate patients on how these activities can be changed into healthy activity levels and

patterns. The psychologists coaches patients in understanding and dealing with the social and

emotional impact of pain in daily life, pain beliefs, and barriers for behavior change, and coach

patients on how to cope with pain. The rehabilitation physicians are responsible for the medical

diagnosis and interventions, and the overall treatment plan. During intake, the rehabilitation

team decides whether the patient can be admitted for PRP. PRP consisted of occupational therapy

sessions, 30 min, two times a week, physiotherapy sessions, 30-60 min, two times a week, and

psychology sessions, 60 min, once a week. The session frequency of occupational therapy and

physiotherapy reduced to once a week during the second half of PRP to encourage patients self

management among patients. After intake, the rehabilitation physician proposed a duration of

8, 12, 16 or 20 weeks of PRP to the patient on the bases of assessment of the complexity of the

physical, social and personal situation of the patient, motivation, and ability to change behavior.

The duration in weeks of the PRP could be adapted (increased or reduced) depending on clinical

progress. Duration could be increased when additional decrease in disability was expected by

lengthening the program. Duration was reduced when treatment goals were achieved earlier

than expected, when the patient showed continued lack of progress, or stops the treatment

because it did not match patients’ expectations or for reasons not related to the program.

Patients were included for PRP when they were 18 years or more of age, had CMP for more

than 3 months, experienced disabilities because of CMP, were not attending any other type

of treatment, were willing to participate in a PRP, and had signed an informed consent form.

Patients with severe comorbidities, interfering with PRP, such as heart failure, rheumatoid arthritis,

psychiatric disorders, and participants with an indication for surgery were not included for the

PRP. All patients treated in the PRP of the UMCG from March 2010 until December 2010 were

eligible for the study.

At baseline, patient characteristics (age, sex, marital status and employment status) were recorded.

The fi rst visit to the PRP (T1) and every uneven treatment week during the program (T2= week 3,

T3= week 5 etc), patients fi lled in the Pain Disability Index (PDI) and a numeric rating scale (NRS)

for average pain in the previous week (0 = no pain, 10= unbearable pain). The last assessment

was performed during the last visit. Patients were blinded for previous scores. If patients could

not be measured at the proposed week, they fi lled in the questionnaire the next visit.

R1R2R3R4R5R6R7R8R9


48 | Chapter 3

Main outcome

The PDI measures self reported disability on seven domains: family/home responsibilities,

recreation, social activity, occupation, sexual behavior, self-care, and life-support activity.

The patient scores the amount of interference in daily life because of pain for each domain

(0= no interference, 10= total interference). Total PDI is calculated by adding the scores of all

domains [range 0 (no disability) to 70 (extreme disability)]. Internal consistency for the PDI is high

(Cronbach’s α=0.86) 10. Construct validity is supported by a multiple regression analyses showing

nine variables to predict the PDI scores (R=0.74) 10. Test-retest reliability (1-week interval) is good

(intraclass correlation coeffi cient=0.91) 11. The minimal clinically important change is 8.5 point 12.

A maximum of two missing items was accepted for the PDI. If more than two items were missing,

a sum score was not calculated. When one or two items of the PDI were missing, the mean score

of the other items of that patient was calculated. The total score was calculated by replacing the

missing item by this calculated mean.

As patients provided written permission to use their clinical data and data were gathered during

care as usual, in the Netherlands, approval from a Medical Ethics Committee is not needed.

Statistical analysis

Descriptive statistics were computed for patient characteristics and outcome variables. Means

and SDs were used for PDI and pain scores. An independent sample t-test was used to compare

means of initial PDI of completers (participants who completed the program) versus noncompleters

(participants who dropped out before the program was ended). Q-Q plots were constructed

to check the assumptions of normal distribution of initial PDI scores. The duration of PRP was

categorized according to the total number of weeks that a patient received treatment. Categories

were: ≤ 8, 9 ≤ 12, 13 to ≤ 16 and ≥17 weeks. To analyze the course of disability during PRP and

factors infl uencing this course, a linear mixed-effect model was applied. A linear mixed-model

analysis corrects for autocorrelation in data, which occurs in repeated measurements over time

within individuals. All available data were taken into account in the analysis, including those of

patients with incomplete datasets 13,14. Outcomes on the PDI were entered as response variable.

Before this analysis, the change in PDI scores during PRP was assessed in a graph to check linearity

for applying the appropriate regression model. Possible predictor variables of the course were

patient characteristics (age, sex, marital status, and employment status), treatment week (=week

of measurement), duration of PRP (in categories), initial PDI score, pain characteristics (average

pain in the week before the PDI assessment, pain duration, and pain localization) and completing

PRP (yes, no). These variables were entered stepwise in the regression model. Assessment of

the graph showed that the change of PDI during PRP allowed a quadratic (treatment week x

treatment week) or cubic model (treatment week x treatment week x treatment week). These

models were explored by entering variable treatment week2 and treatment week3 in the regression



3

equation. Entered predictors remained in the model if the model fi t (-log-likelihood) increased

signifi cantly or if the regression coeffi cient was signifi cant. After the main effects were established

the covariance structure best applicable to the data was explored, again using the model fi t as

criterion. After determining the covariance structure, random intercept and slopes were explored

using the model fi t. Interactions between predictors were explored if main effects were signifi cant

(P≤0.05). Data were analyzed using SPSS (version 18.0; SPSS Inc., Chicago, Illinois, USA).

Results

A total of 128 patients participated in the study. Mean age of the patients was 41.7 (SD 11.8,

range 19-68) years. CMP was categorized as chronic back pain (55%), chronic neck pain

(19%) and other types of chronic pain (26%), such as widespread pain, non-specifi c pain of

lower extremities (knee, foot) etc. Patients attended PRP for an average duration of 12.5 (SD

4.0, range 3-23) weeks and received on average 38 (SD 17.1, range 5.5-89.8) contact hours.

Thirty-eight percent (n=49) of the patients received the proposed amount of treatment weeks

(range ± 1 week). Thirty-seven percent (n=47) received less (including noncompleters) and 13%

(n=17) received more than amount of treatment weeks proposed by the rehabilitation physician.

The proposed duration was not reported in 12% (n=15) of the cases. Twenty-six (20%) of the

128 patients dropped out during the PRP; they did not complete the program for lack of effect

(2%, n=2), unrealistic expectations of the program (3%, n=4), wanting additional diagnostic

procedures (2%, n=2), because they believed that the benefi ts of the program did not weigh

against efforts (3%, n=4), treatment was too demanding (3%, n=4) or other reasons (8%,

n=10). The majority of noncompleters stopped during the fi rst 8 weeks of PRP. Because of

different treatment durations and dropout, the total number of participants decreased from the

fi rst to the 27th treatment week (Fig. 2). Baseline characteristics (n= 128) are shown in Table 1.

Noncompleters and completers signifi cantly differed on age, marital status, and employment

status. The difference in duration of pain of completers versus noncompleters was not signifi cant.

The mean PDI for all patients at T1 (n = 125) was 34.2 (SD 12.0). Mean NRS average pain was 5.6

(SD 1.9). The mean initial PDI was signifi cantly (P=0.013) lower for completers compared to

noncompleters [mean difference= 6.6, confi dence interval (CI) 1.4-11.7] (Table 2).

R1R2R3R4R5R6R7R8R9


50 | Chapter 3

Tab

le 1

Bas

elin

e ch

arac

teris

tics

(n=

128)

All

patie

nts/

du

ratio

nsSo

rted

by

dura

tion

PRP

(wee

ks)

Com

plet

ers

vers

us n

onco

mpl

eter

s

≤ 8

9 ≤

12

13 ≤

16

≥17

Com

plet

ers

Non

com

plet

ers

P va

lue*

(n=

128)

(n=

21)

(n=

40)

(n=

52)

(n=

15)

(n=

102

)(n

= 2

6)A

ge [m

ean

(SD

)]41

.7 (1

1.8)

45.7

(12.

3)40

.5 (1

3.3)

42.0

(10.

2)37

.9 (1

1.7)

40.6

(11.

7)45

.8(1

1.6)

0.04

5Se

x (f

emal

e) [n

(%)]

73 (9

4)91

(19)

88 (3

5)56

(29)

73 (1

1)71

(72)

85 (2

2)0.

148

Mar

ried

[n (%

)]73

(94)

95 (2

0)70

(28)

75(3

9)47

(7)

69 (7

0)92

(24)

0.02

0M

issi

ng

2 (2

)0

(0)

0 (0

) 4

(2)

0 (0

)2

(2)

0 (0

)Em

ploy

men

t st

atus

[n (%

)]0.

019

Hou

sem

an/-

wife

8 (1

0)24

(5)

2 (1

)6

(3)

7 (1

)5

(5)

19 (5

)A

tten

ding

edu

catio

n6

(8)

0 (0

)10

(4)

6 (3

)7

(1)

8 (8

)0

(0)

Wor

king

or

sick

leav

e73

(93)

62 (1

3)73

(29)

75 (3

9)79

(12)

76 (7

7)62

(16)

Dis

abili

ty c

ompe

nsat

ion

11 (1

4)14

(3)

13 (5

)9

(5)

7 (1

)9

(9)

19 (5

)M

issi

ng2

(3)

0 (0

)2

(1)

4 (2

)0

(0)

3 (3

)0

(0)

Loca

lizat

ion

of p

ain

[n (%

)]0.

836

Low

bac

k pa

in55

(70)

53 (1

1)50

(20)

60 (3

1)53

(8)

54 (5

5)58

(15)

Nec

k pa

in19

(25)

14 (3

)23

(9)

17 (9

)27

(4)

21 (2

1)15

(4)

Oth

er26

(33)

33 (7

)27

(11)

23 (1

2)20

(3)

26 (2

6)27

(7)

Dur

atio

n of

pai

n [m

edia

n (IQ

R)] (

year

s)

3.0

(1.5

-8.0

)

5.0

(2.0

-15.

0)3.

0 (1

.5-1

0.0)

3.0

(1.5

-7.0

)2.

8 (2

.1-4

.5)

3.0

(1.5

-7.0

)

4.5

(2.0

-18.

8)

0.20

8

Mis

sing

[n (%

)]22

(28)

10 (2

)28

(11)

23 (1

2)20

(3)

26 (2

6)8

(2)

PRP,

pai

n re

habi

litat

ion

prog

ram

.*

P va

lue

of d

iffer

ence

bet

wee

n co

mpl

eter

s an

d no

ncom

plet

ers.



3

Course of outcome Pain Disability Index scores over time

The mean PDI scores and the mean average pain scores on the previous week over time are

shown in Fig. 2. The wide 95% CI at the even treatment weeks and at the tail of the curve (>17

weeks) can be attributed by the smaller number of patients measured in those weeks (n= shown

under the x-axis). From week 17, the number of participating patients decreases to less than 1/3

of the total number of patients at week 1; therefore the 95% CI’s become wider and the scores

are less precise.

Table 2 Mean PDI and average pain at T1.

Variable Total groupn=125Mean (SD)

Completersn= 99Mean (SD)

Non-completersn=26Mean (SD)

Differences in means (95% CI)

p value

PDI (0-70) 34.2 (12.0) 32.8 (11.8) 39.4 (11.4) 6.6(1.4 to 11.7) 0.013Average pain (0-10) 5.6 (1.9) 5.5 (1.9) 5.7 (1.8) 0.2 (-0.6 to 1.0) 0.627

119* 6 107 14 93 22 83 23 83 21 58 30 38 18 33 28 15 3 8 5 4 3 2 1 2-1

0

1

2

3

4

5

6

7

8

9

10

-10

0

10

20

30

40

50

60

70

80

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 25 27

Mea

n o

f av

erag

e p

ain

in p

revi

ou

s w

eek

Mea

n P

DI a

nd

95%

CI PDI

Pain

Figure 2 Mean PDI and 95% CI and mean of average pain.n= Number of available measurements reported per week.

Mixed-effect modeling

The covariance structure was autoregressive heterogeneous, which means that correlations

between scores reduced over time 13. Variables that were entered in the model but did not

improve the model signifi cantly were patient characteristics (age, sex, marital status, and

employment status), duration of the PRP (≤8, 9 to ≤12, 13 to ≤16, and ≥17 weeks), pain duration,

pain localization and interaction between duration of PRP and treatment week. This means that

these variables did not infl uence the course of reduction on PDI over time. Predictor variables that

R1R2R3R4R5R6R7R8R9


52 | Chapter 3

signifi cantly infl uenced the model were treatment week, initial PDI score, average pain, and the

interaction between initial PDI and treatment week. This means that these variables did infl uence

the course. The model improved signifi cantly by allowing for each patient a random intercept

and a random slope for average pain (Table 3). Treatment week2 contributed signifi cantly toward

the regression equation, treatment week3 did not. This means that the PDI reduces according to

a quadratic model over time during PRP, and not according to a cubic model.

Table 3 Results of a linear mixed-model analysis to predict change in Pain Disability Index score during a pain rehabilitation program.

Predictor βa SE β P valueIntercept (Pb) -5.3 1.9 0.005Treatment week -0.2 0.2 0.174Initial PDI 0.8 0.05 < 0.001Average pain previous week (Pb) 2.3 0.2 < 0.001Time X Initial PDI -0.02 0.004 < 0.001Treatment week x Treatment week 0.03 0.007 < 0.001

PDI, Pain Disability Index.aβ: Regression coeffi cient.bP: random for patient.

Clinically, this means that treatment week, initial PDI score, average pain, and the interaction

between initial PDI and treatment week infl uence the course of PDI scores during PRP. The

coeffi cients of Table 3 indicate that, for instance, a patient with an initial PDI score of 40 and a

score of 6 on average pain in the fourth treatment week will have an average PDI score of 36.3 in

the fi fth treatment week of the PRP. This can be calculated using the following formula:

PDI= -5.3 + (-0.2 x treatment week) + (0.8 x initial PDI) + (2.3 x average pain) + (0.03 x treatment

week2) + [-0.02 x (initial PDI x treatment week)].

Discussion

At group level, disability reduction during PRP occurs following a quadratic model, quite similar

to the hypothesized line B in Fig. 1. Reduction in average pain in the previous week precedes

decrease of disability. The strength of this infl uence is 2.3 points on PDI per point pain reduction

(Table 3). Hence, a reduction of 1 point on average pain will decrease disability with 2.3 points.

The interaction between initial PDI and treatment weeks also infl uences the course, which means

the strength of infl uence per treatment week on PDI (β=-0.02) depends on the initial PDI: a

higher initial PDI will lead to a steeper decrease per week compared with a lower initial PDI. The

infl uence of the quadratic factor of treatment week on the course is 0.03: the effect of treatment

week on PDI (-0.2 per treatment week) is reduced by the quadratic term of treatment week



3

(0.03). This means that towards the end of the PRP, disability reduction per week becomes smaller

(Fig. 1, line B).

The results of this study have provided insight into the course of disability reduction over time. The

clinical relevance of this study is two-fold. First, because most of the gain in disability reduction

is obtained in the earlier weeks of PRP, the added value of additional PRP toward the end of

the program may become trivial. Second, the relevance of knowing which factors infl uence the

course of disability reduction may be that one can on average predict the duration of the PRP

if one knows the target reduction of disability. For example, if the target reduction is 8.5 PDI

points, the minimal clinically important change 12, the predicted mean duration of the PRP for an

‘average patient’, with an initial PDI of 40 and an average pain score of 6, would be 14 weeks.

This is the fi rst study in which the course of disability during a PRP was analyzed. Most studies

have focused on the effect of PRPs based on pre, post, and follow-up measurements, focused on

other types of treatments or other groups of participants 6,15,16. With this study, we have provided

a new aspect of research of PRPs: dose of PRP. Within the rehabilitation literature, there is no

robust evidence to substantiate dose aspects (duration and contact hours) of PRPs. It appears that

dose aspects have been neglected in PRP research. Generalization of the fi ndings of this study may

be challenging and calls for replication studies within and outside the Netherlands. The general

principle of the course of disability reduction might also apply to PRPs outside the Netherlands.

However, dose choices are also infl uenced by cultural factors and differences between professional

teams within and outside the Netherlands. In addition, fi nancial reimbursement will also infl uence

the choice of dose.

Study limitations

This study has some limitations. Some data were missing on PDI items. However, the effects

of these missing data were limited because the total score of PDI is used, which represents

general (average) disability. For the statistical analyses, the use of a mixed model limited the

infl uence of missing data. In clinical practice, we could not arrange a rigidly fi xed measurement

time equal for each patient. We had to deal with absence of patients and care as usual was

not ready for digitalized versions of the questionnaires to minimize missing data because of

absence. Nevertheless, all available data were used in the mixed-effect model. Because of ethical

regulations related to digital patient information safety, we could not follow up patients every

2 weeks after PRP. It is advised, however, that future studies focus on the follow up period also

because it is assumed that further progress occurs after completion of the program. Twenty

percent of the patients dropped out, but data of all patients, completers and noncompleters,

were taken into account for the mixed-model analysis. Noncompleters may be considered a

limitation from a statistical point of view. However, in daily clinical care, noncompletion is often

R1R2R3R4R5R6R7R8R9


54 | Chapter 3

observed. Noncompletion has not been the subject of study. In this study, it appeared challenging

for the clinicians to estimate treatment duration before the PRP. As shown in the result section,

the actual PRP duration differed from the pretreatment estimated duration. This may be regarded

a limitation of this study because this provides variance in total treatment weeks and number

of patients per treatment duration, but it underscores the clinical need to know the course of

disability reduction and factors predicting this course that could contribute toward more rational

estimations of treatment duration.

The major strength of this study is the strong relation to clinical practice, which enhances the

external validity: measurements were performed in care as usual and the results of the mixed-

model analyses can contribute toward estimate of the duration of PRPs. Also, the heterogeneous

group strengthens the results of this study because differences in pain locations did not affect

the results and therefore the outcome is generalizable to a broader group of patients with CMP.

Conclusions regarding optimum duration of PRPs cannot be drawn. However, the results of this

study show the need for further research on the course of disability and dose aspects of PRPs.

Future research could focus on the effect of shortening a PRP on the course of disability reduction

or the course of disability after completion of a PRP. It should also focus on the added value of

additional PRP (h/weeks) to analyze at which point treatments are no longer benefi cial for the

patient or no longer cost-effective.

Conclusion

From the results of this study, we can conclude that disability reduces according to a quadratic

model during PRPs. Initial PDI, treatment week, average pain, and interaction between initial PDI

and treatment week infl uence the course of disability reduction during PRP. These data could aid

in clinical practice to predict the duration of the PRP at start. Further research investigating dose

aspects of PRPs is needed to aid in the effectiveness and cost-effectiveness of PRPs.

Acknowledgements

Many thanks are due to all the therapists of the Pain Rehabilitation Department of the Center for

Rehabilitation of the University Medical Center Groningen for collecting the data.



3

References

1. McCracken LM, Turk DC. Behavioral and cognitive-behavioral treatment for chronic pain: Outcome, predictors of outcome, and treatment process. Spine (Phila Pa 1976). 2002;27(22):2564-2573.

2. Gatchel RJ, Okifuji A. Evidence-based scientifi c data documenting the treatment and cost-effectiveness of comprehensive pain programs for chronic nonmalignant pain. J Pain. 2006;7(11):779-793.




6. Skouen JS, Grasdal AL, Haldorsen EM, Ursin H. Relative cost-effectiveness of extensive and light multidisciplinary treatment programs versus treatment as usual for patients with chronic low back pain on long-term sick leave: Randomized controlled study. Spine (Phila Pa 1976). 2002;27(9):901-9; discussion 909-910.

7. Jensen IB, Bergstrom G, Ljungquist T, Bodin L, Nygren AL. A randomized controlled component analysis of a behavioral medicine rehabilitation program for chronic spinal pain: Are the effects dependent on gender? Pain. 2001;91(1-2):65-78.


9. van Hooff ML, van der Merwe JD, O’Dowd J, et al. Daily functioning and self-management in patients with chronic low back pain after an intensive cognitive behavioral programme for pain management. Eur Spine J. 2010;19(9):1517-1526.

10. Tait RC, Chibnall JT, Krause S. The pain disability index: Psychometric properties. Pain. 1990;40(2):171-182.

11. Gronblad M, Hupli M, Wennerstrand P, et al. Intercorrelation and test-retest reliability of the pain disability index (PDI) and the oswestry disability questionnaire (ODQ) and their correlation with pain intensity in low back pain patients. Clin J Pain. 1993;9(3):189-195.

12. Soer R, Reneman MF, Vroomen PC, Stegeman P, Coppes MH. Responsiveness and minimal clinically important change of the pain disability index in patients with chronic back pain. Spine (Phila Pa 1976). 2012;37(8):711-715.

13. Field A. Multilevel linear models. In: Discovering statistics using SPSS. 3rd ed. Los Angeles, London, New Delhi, Singapore, Washington DC: Sage; 2009:725.

14. Twisk JWR. Applied longitudinal data analysis for epidemiology: A practical guide. Cambridge etc.: Cambridge University Press; 2003:301.

15. Gross DP, Haws C, Niemelainen R. What is the rate of functional improvement during occupational rehabilitation in workers’ compensation claimants? J Occup Rehabil. 2012;22(3):292-300.

16. Rainville J, Jouve CA, Hartigan C, Martinez E, Hipona M. Comparison of short- and long-term outcomes for aggressive spine rehabilitation delivered two versus three times per week. Spine J. 2002;2(6):402-407.

FrFrFrananankakaka PPP.CCC. WaWaWatetetersrsrschchchoooooot,t,, PPieieeteteter rr U.U.U DDDijiji ksksk trtraaa PhPhPhD,D,DJoJoJohahahannnn esses GGG.MM.M.. BuBuBurgrggerererhohoof f PhPhPhD,D,D JJananan HHH.BBB. GeGeGeerere tzztzenen MMD,D,D, PPPhDhDhD,

HeHeHenrnrn icicicaaa R.RR SSchchipipphohohorsst tt PrPrPreueue pepeper r r MDMDMD,,, PhPhP D,D,D, MMMiccchihhieleel FFF. ReReR nennemamam nnn PhPhPhD.DD

SuSuSubmbmbmitittteteted dd fofoor r pupublblblicicicatatatioioonn

R1R2R3R4R5R6R7R8R9


58 | Chapter 4

Abstract

Objective: To analyze effects of multidisciplinary pain rehabilitation programs with different

dosages; care as usual (CAU) versus care as usual short form (CAU-SF).

Methods: A single blinded, 2 armed, randomized controlled trial, with non-inferiority design

was performed. All patients with chronic musculoskeletal pain referred to an outpatient

multidisciplinary pain rehabilitation program were eligible for this study. Only dosage in weeks

differed between the 2 groups, content was similar. The pain disability index (PDI) was the primary

outcome measure. Four points difference on PDI was applied as non-inferiority margin. Treatment

effects within groups were expressed in standardized mean difference (SMD) and effect sizes (ES)

were calculated between groups.

Results: Both groups improved signifi cantly on PDI (CAU: -10.8, CAU-SF: -8.3). Frequent extension

of CAU-SF leads to limited contrast in dosage between groups. The 2.5 points difference on PDI

falls within the non-inferiority margin but the confi dence interval (CI) (-2.2 to 7.3) exceeded it.

SMDs of CAU and CAU-SF were 0.8 and 0.7 respectively. ES was 0.2.

Conclusions: Reduction of dosage of pain rehabilitation programs did not lead to non-inferior

mean results. Difference in means showed non-inferiority but CI exceeded both the upper and

lower border of the non-inferiority margin. The results of this trial are inconclusive.

Keywords

Dosage, Pain Rehabilitation Program, Chronic pain, Chronic musculoskeletal pain.


Dosage of PRP: A RCT | 59

4

Introduction

Multidisciplinary Pain Rehabilitation Programs (PRPs) are recommended for patients with chronic

low back pain (CLBP) 1. They aim to reduce disability, distress and use of health care services

by education of physical, psychological and practical techniques to improve function, work

participation and health related quality of life 2. Multidisciplinary programs are effective for

patients with chronic musculoskeletal pain (CMP) 1,3,4.

Across pain rehabilitation centres nationally and internationally, there is much diversity in content

and dosage of PRPs. Health care systems aim for the best treatments based on available evidence.

However, there is paucity of evidence about the infl uence of dose on effects of PRP. In a recent

systematic review 5, no randomized controlled trials (RCTs) were identifi ed that were designed to

analyse effects of differences in dose variables on outcome of PRPs. Additionally no studies were

identifi ed with the primary objective to analyze the association between dose and effect of PRP.

If PRPs optimal dosage is known, this may benefi t patients and could reduce direct or indirect

costs. If similar effects are achieved with a shorter program, this could lead to earlier reduction of

disability, regaining quality of life, and sooner participation in daily life, and minimize direct and

indirect costs associated with PRP. Employers could also benefi t from earlier return to work (RTW)

for patients with work productivity loss, rehabilitation centres can reduce waiting lists and overall

it can improve effi ciency of care. Additionally shorter programs are also attractive for health care

insurers and society as a result of reducing direct and indirect costs.

The aim of this study was to analyze differences in effects of PRPs with different duration: care as

usual (CAU) and PRP short form care as usual (CAU-SF) in a non-inferiority design. CAU-SF was

hypothesized to be non-inferior to CAU in an outpatient pain rehabilitation setting.

Methods

Design

The study is a single blinded, 2 armed, RCT, with a non-inferiority design. The RCT took place

at the University Medical Centre Groningen (UMCG), Department of Rehabilitation Medicine,

Centre for Rehabilitation, location Beatrixoord Haren. The Medical Ethics Committee of the

UMCG approved the study protocol [NL30094.042.11]. Trial registration: international clinical

trials registry platform of WHO, TrialID= NTR3385.

R1R2R3R4R5R6R7R8R9


60 | Chapter 4

Defi nitions

Because dosage variables are relatively new to the scientifi c fi eld of PRP, we have defi ned dosage

related terms for this study. Duration means the total length of PRP and is expressed in weeks.

Contact hours are the total amount of hours that a patient spends with his therapist during PRP.

Participants

All patients with CMP referred to an outpatient PRP at the Centre for Rehabilitation, UMCG,

between September 2011 and October 2013 were potential participants for this study. This

centre provides PRPs with duration of 8, 12, 16 or 20 weeks. Patients were eligible for the study

when: 1. They were admitted for a 12, 16 or 20 week PRP; 2. they had CMP for more than three

months without a specifi c pathological cause; 3. they experienced CMP induced disability; 4.

social and psychological factors were complex and were assumed to be relevant in maintaining

CMP induced disability 2; 5. they were willing to stop other treatments for CMP during PRP (except

pain medication); 6. they were 18 years or older; 7. they were motivated to participate in PRP

and 8. they were willing to participate in the study and signed informed consent. Patients were

excluded when: 1. they were referred to the 8 weeks PRP; 2. they were unable to understand

the Dutch language; 3. they had co morbidities such as heart failure, rheumatoid arthritis, or

psychiatric disorders preventing a PRP.

Interventions

Common features of CAU and CAU-SF

The objectives of CAU and CAU-SF were the same and the content was similar; both treatments

were outpatient, multidisciplinary PRPs aimed to decrease CMP related disability, optimize

participation and increase quality of life. PRP is intended to coach patients to self manage their

pain and disability. The rehabilitation team consisted of rehabilitation physicians, occupational

therapists (OTs), physiotherapists (PTs) and psychologists (PSYs). The PRP, based on cognitive

behavioural principles, consisted of pain education, for example regarding differences between

nociceptive and chronic pain. Patients were counselled to refl ect on their pain management

strategies (avoiding pain) and how these strategies could be changed into other management

strategies (pain coping, alternating physical activity and rest and gradually increasing activities).

Each patient set individual treatment goals. PTs applied exercise programs and sports activities,

to improve patients’ confi dence in movement and reduce pain-related fear, improve activity level,

and physical functioning. OTs assessed current activities and patterns in daily living and educated

patients how activities could be changed into, healthy activity levels and patterns. PSYs coached

patients in understanding and dealing with the social and emotional impact of pain in daily life,

pain beliefs, barriers for behaviour change and coached patients how to cope with pain. The

rehabilitation physicians were responsible for the medical diagnosis and interventions (if any) and

coordinated the overall treatment plan.



4

After the rehabilitation physician considered patients to be eligible for PRP, they were invited for

an intake procedure, consisting of an interview by an OT, a PT and a PSY. These interviews were

aimed at admitting a patient to PRP or not, and to formulate a treatment plan. If the patient was

admitted to PRP, based on the assessments during intake, the team (rehabilitation physician, OT,

PT, PSY) determined the required duration of 8, 12, 16 or 20 weeks based on assessment of the

complexity of the physical, social and personal situation of the patient, motivation and ability to

change behaviour. The team who determined treatment duration consisted of professionals with

on average 9 years of experience in pain rehabilitation.

PRP consisted of OT sessions, 30 minutes, 2 times a week, PT sessions, 30 to 60 minutes, 2

times a week and PSY sessions, 60 minutes, once a week. The session frequency of OT and

PT reduced to once a week, the PSY sessions reduced to once every 2 weeks, based on the

frequency protocols per duration. All sessions were delivered in a individual setting. For clinical

reasons, the duration in weeks of PRP could be adapted (extended or shortened) in CAU and

in CAU-SF, depending on the progression of the patient. Duration could be extended when

additional coaching was needed to decrease in disability during the extra weeks of PRP. Duration

was shortened when treatment goals were achieved earlier than expected, when the patient

demonstrated continued lack of progress, or when the patient stopped the treatment, because it

did not match patient expectations or reasons not related to the program (for instance holidays).

Extending or shortening of the duration of PRP was based on agreement between the patient and

the PRP team and decided by the rehabilitation physician. Reasons for extending or shortening

PRP were registered.

The difference between CAU and CAU-SF was the duration (in weeks) of PRP proposed to the

patients after the randomization procedure. Differences in duration will also result in differences

in contact hours. Patients allocated to the CAU group received 12, 16 or 20 weeks of PRP as

proposed after the intake procedure. Patients allocated to the CAU-SF group received a 4 week

shortened form of PRP than proposed (which will result in approximately 10 fewer contact hours

on average). Thus patients in CAU-SF received 4 weeks less than proposed after intake. Patients

allocated to 8 weeks of PRP in CAU were excluded because the treatment team thought it would

not be feasible to offer PRP in 4 weeks with frequencies twice a week.

Outcomes

The primary outcome variable was self-reported disability assessed with the Pain Disability Index

(PDI) 6. The questionnaire consists of seven items related to work, leisure and activities of daily life

each item is scored on an 11-point scale (0 indicating no disability and 10 indicating maximum

disability). The total scale ranges from 0 to 70. The high correlation between the PDI and Oswestry

Disability Index (r=0.83) indicates good validity. Test-retest reliability is high (ICC=0.91) 7. Minimal

Clinically Important Change (MCIC) ranges from 8.5 to 9.5 points 8.

R1R2R3R4R5R6R7R8R9


62 | Chapter 4

The secondary outcome variable, quality of life, was measured with the Euroqol 5D-5L (EQ5D-5L;

scale range -0.329 to 1.000) 9. The EQ-5D-5L assesses 5 dimensions (mobility, self-care, usual

activities, pain/discomfort, anxiety/depression) of health on a 5 point scale: no problems, slight

problems, moderate problems, severe problems, and extreme problems. The overall health is

assessed on a 0-100 visual analogue scale of 20 cm (EQ-5D-VAS). A higher score indicates a

better overall health state. The minimally clinical important difference (MCIC) was 10.5 points for

the EQ-5D VAS and 0.03 points for the 5 dimensions of the EQ-5D-3L 10. The EQ-5D-5L version is

based on the EQ-5D-3L version. The 5L version demonstrates improved discriminatory power 11,

but the exact MCIC for this version has not been examined (yet).

Sample size

Sample size was calculated using the group sequential non-inferiority criterion. The standard

deviation (SD) on PDI scores in CAU was 11. The non-inferiority margin was calculated as 40%

of the mean change on PDI scores in CAU, which was 10 points, thus the non-inferiority margin

was set at 4 points. For group sequential non-inferiority design an alpha of 0.045 and power of

0.8 was used. Using PASS software, these data gave an estimation of 124 patients needed per

arm and 248 patients in total. Including 10% drop-outs, 276 patients should be included for the

study.

Randomization

After determination of PRP duration by the team and after obtainment of informed consent,

patients were randomized by an independent person. Patients were stratifi ed on work status

(paid work or unpaid/no work), resulting in 2 strata. For each stratum blocks of 6 were used.

Sequentially numbered, sealed, opaque envelopes were used for each stratum.

Blinding

Patients were blinded for intervention. Before inclusion, they were informed about duration

of CAU, which could be 8, 12, 16 or 20 weeks. Patients were explained that a duration was

proposed by the team based on the assessment results. Furthermore, they were informed about

the experiment of 4 weeks shortening duration. They were unaware of the treatment duration

they supposed to receive in CAU, and, therefore, they were blinded of allocation to CAU or CAU-

SF. Patients were outcome assessor for outcome measurements; they were blinded for T0 scores.

Rehabilitation physicians, OTs, PTs and PSYs could not be blinded for intervention for the 20

weeks duration because this part only exists in the CAU group. Eight week PRPs could be CAU-SF

or patients not included in the study attending 8 weeks PRP.



4

Analyses

An interim analysis was performed halfway the inclusion period. This enabled us to recalculate

the a priori sample size based on trial data, as opposed to assumptions (a priori). Additionally,

based on a-priori stopping rules, it was decided that the trial would be stopped if more than 25%

of the PRPs in the CAU-SF were extended or if the difference in means on PDI between CAU and

CAU-SF was more than 4 points. Differences in mean PDI scores were tested by one sided t-test.

In consultation with a statistician, a priori it was decided to split the alpha of 0.05 in 0.005 and

calculate a two-sided 99% confi dence interval (CI) for the interim analysis, and 0.045 with a two-

sided 91% CI for fi nal analyses. Because the study was stopped preliminarily, however, only one

analysis was performed for which we used a two-sided 95% CI approach.

Means and SD’s of primary outcome are presented for CAU and CAU-SF group. All statistical

analyses were performed according to an intention to treat analysis and a per protocol analysis.

A CI approach was used to interpret non-inferiority (Figure 1). For PDI and EQ5D scores, 95%

2-sided CIs were used in the fi nal analyses. Non-inferiority was established when the difference in

means was less than 4 (Figure 1). Treatment effects within groups were expressed in standardized

mean difference (SMD): (meanchange/ SDchange). Effect sizes (ES) were calculated between groups

(meanchange intervention group - meanchange control group)/ SDPooled 12. Interpretation of SMD and ES: <0.10 no

effect; <0.30 little effect; <0.50 moderate effect; <0.80 large effect; >0.80 very large effect 13.

0 4

D

C2

B3

B2

B1

A

Superior Equivalent Non Inferior Inferior

2.53 -2.2 7.3

C1

Figure 1 Illustration of the phenomenon of non-inferiority testingThe dotted horizontal lines represent the theoretical width of the Confi dence Intervals (CI) of the differences in means between experimental group and control group. If the CI lies left of 0: (A) the experimental group is superior to the control group. If CI lies left of 4 (B1, B2, B3) the experimental group is non inferior. If CI includes 4 (C2) non inferiority is inconclusive. If CI lies is right to 4 the experimental group is inferior (D). The statistical analyses of the differences between groups resulted in line C1 with confi dence intervals for -2.2 to 7.3: which means non inferiority is inconclusive

R1R2R3R4R5R6R7R8R9


64 | Chapter 4

Results

Patient characteristics and trial fl ow

Patients were recruited out of an estimated sample of 694 patients, admitted for PRP between

September 2011 and October 2013. In total 257 (37%) were eligible after signing informed

consent. Reasons for non-participation were for example: planning misunderstandings, trial

participation was of low priority and thus forgotten to sign informed consent, unwilling to

participate in a trial. Interim analyses (October 2013) revealed that in 36% of the patients in

the CAU-SF PRP was extended. Based on the a priori stopping rules, the study was stopped

directly after the interim analyses. At that point, n=201 (102 CAU, 99 CAU-SF) patients were

randomized (Figure 2). In accordance with the rules of the METC, 30 patients, who were enrolled

in the trial, were given the possibility to receive the admitted treatment duration and therefore

these patients were excluded from analyses. Analyses were performed for 153 patients who had

completed PRP: 81 in CAU and 72 in CAU-SF. The majority of patients experienced pain for more

than 1 years. The most prevalent diagnoses were chronic back pain (42%), chronic neck pain

(19%), widespread pain (9%) and fi bromyalgia (7%). Patients who completed the study differed

signifi cantly on work status and welfare status, compared to those who did not complete the

study (Table 1).

Assessed for eligibility (n=257)

Excluded (n=56) Not meeting inclusion criteria (n=43) Excluded because stop RCT (n=13)

Randomized (n=201)

Analysed (n= 81) Analysed (n=72)

Allocated to CAU-SF (n=99) Received allocated intervention (n=72) Did not receive allocated intervention (n=12) Excluded because stop RCT (n=14) Declined to participate (n= 1)

Allocated to CAU (n=102) Received allocated intervention (n= 81) Did not receive allocated intervention (n= 5) Excluded because stop RCT (n=16)

Enrollment

Allocation

Analysis

••

••

•

••••

Figure 2 Flow diagramCAU: Care as usual, CAU-SF: Care as usual short form



4

Table 1 General characteristics of participants in the CAU and the CAU SF and participants who completed the study and those who did not.

CAUn=81

CAU-SFn=72

Completersn=120

Drop-outsn=33

mean SD mean SD median IQR median IQRAge (years) (n=153) 44 12.4 44 11.9 44.6 33.6 to 52.3 47.4 40.4 to 54.9Average pain last week (n=153) 6 1.6 7 1.6 7 6 to 8 7 5 to 8Worst pain last week* (n=153) 8 1.6 8 1.3 8 7 to 9 8 7 to 9

n % n % n % n %Gender (% female) (n=153) 45 56 49 68 79 66 15 45Duration of pain (n=131)

3 months to 1 year 15 23 20 30 27 26 8 31> 1 year 50 77 46 70 78 74 18 69

Assigned treatment duration (weeks)(n=153)

8 0 0 26 36 23 19 3 912 37 46 45 63 66 55 16 4916 41 51 1 1 30 25 12 3620 3 4 0 0 1 1 2 6

Marital status (n= 132)Single 9 14 11 17 16 15 4 15Married/ living together/LAT 53 80 47 71 77 73 23 85Otherwise 4 6 8 12 12 12 0 0

Education (n=136)No or low education 20 30 13 19 25 23 8 30Middle education 30 44 34 50 51 47 13 48High education 18 17 21 31 33 30 6 22

Work status** (n=135)Employed, fulltime 19 27 20 31 34 32 5 18Employed, parttime 22 32 18 27 36 34 4 14houseman/-wife 9 13 8 12 9 8 8 29Unemployed 8 12 8 12 12 11 4 14Otherwise 11 16 12 18 16 15 7 25

Welfare status** (n=135)Not working and no welfare 25 37 23 34 43 40 5 18Partial sick leave/disability pension 30 44 30 45 40 37 20 71Full sick leave/disability pension 6 9 9 13 14 13 1 4unemployed 4 6 0 0 4 4 0 0Otherwise 3 4 5 7 6 6 2 7

Pain location (n=333)Head or face 16 9 21 13 31 12 6 8Upper extremities 21 12 16 10 27 10 10 14Neck or shoulders 38 22 39 24 59 23 18 25Low back 47 28 45 27 75 29 17 24Lower extremities 32 19 33 20 50 19 15 21Otherwise 15 9 10 6 20 8 5 7

Number of sites (median, IQR) 2 [1,4] 2 [1,4] 2 [1,4] 3 [1,3]

CAU: Care as usual, CAU-SF: Care as usual short form * Signifi cant difference between CAU and CAU-SF (<0.05) **Signifi cant difference between completers and drop-outs (<0.05)

R1R2R3R4R5R6R7R8R9


66 | Chapter 4

Outcomes and estimation

The reduction of PDI scores was 2.5 larger in the CAU group compared to the CAU-SF group

(Figure 1). The point estimate of the difference was 2.5 and lies within the non-inferiority margin

from 0.0 to 4.0. Because the CI of this difference exceeded both the upper and the lower border

of the non-inferiority margin, the results were inconclusive.

In 74% of the cases, PRPs were delivered per protocol or shorter, and in 26% of the cases PRPs

were extended (Table 2: CAU-SF 36%; CAU 17%). Differences in dosage between shortened,

dosage as planned and extended PRPs were non-signifi cant (p=0.066). Reasons for extending

were: changes in the behaviour or situation of the patient, RTW assistance needed more support

and/or more time than anticipated, patient case was more complex than estimated a priory, and

treatment logistics.

Both groups improved signifi cantly over time. SMD and ES for PDI and EQ5D scores are presented

in Table 3. Differences in results were not signifi cant (p>0.05) between groups for both outcome

measures.

Table 2 Treatment duration

CAUn=81

CAU-SFn=72

P value 95% CI

Received treatment duration (weeks) (mean sd)

11.7 4.5 10.8 3.9 0.323 -0.525 to 2.219

Received contact hours (mean, sd) 30.7 11.3 29.8 10.4 0.622 -2.651 to 4.389PRP Dropouts (n,%) 22 27% 11 15%

Weeks (mean, sd) 6.0 3.7 5.1 2.2PRP Shortened (n,%) 23 29% 15 21%

Weeks (mean, sd) -5.7 4.0 -3.4 2.6PRP As planned (n,%) 22 27% 20 28%PRP Extended 14 17% 26 36%

Weeks (mean, sd) 2.9 2.0 4.0 2.0

CAU: Care as usual, CAU-SF: Care as usual short form, PRP: Pain Rehabilitation Program.



4

Table 3 Trial results (pain disability and quality of life) between and within treatment arms

n T0Mean (sd)

n T1Mean (sd)

n Difference in means (sd)

95% CI SMD

PDI total 144 37.0 (13.3) 119 25.8 (16.2) 112 9.6 (12.8) 7.2 to 12.0 0.8CAU 74 36.1 (12.5) 63 25.1 (15.0) 58 10.8 (13.2) 7.3 to 14.3 0.8CAU-SF 70 37.9 (14.2) 56 26.6 (17.7) 54 8.3 (12.2) 5.0 to 11.6 0.7

ES for between group difference

0.20

EQ5D index total 125 0.55 (0.20) 101 0.70 (0.18) 93 0.13 (0.17) 0.09 to 0.17 0.8CAU 66 0.56 (0.19) 54 0.70 (0.17) 51 0.11 (0.19) 0.06 to 0.17 0.6CAU-SF 59 0.54 (0.21) 47 0.70 (0.20) 42 0.15 (0.15) 0.10 to 0.19 1.0

ES for between group difference

-0.19

T0: before pain rehabilitation program (PRP); T1: direct after PRP.PDI: Pain Disability Index (scale 0-70); EQ5D: Euroqol 5D index (scale 0.0-1.0),CAU: Care as usual, CAU-SF: Care as usual short form, SMD: Standardized mean difference; ES: Effect Size

Ancillary analyses

Further analyses were performed on differences in trial results between subgroups, patients for

whom the PRP was delivered per protocol or shortened, and patients for whom the PRP was

extended. The mean duration of the per protocol subgroup was 10.0 (SD 3.8) weeks. The mean

number of contact hours was 27.1 (SD 9.7). For the extended subgroup mean duration was

15.0 (SD 3.3) weeks, and 39.2 (SD 8.9) contact hours (Table 4). SMD of PDI of the per protocol

subgroup was 0.9; ES of the extended subgroup was 0.6. Because the duration of the PRPs

was similar between groups, a regression analysis was performed to analyze predictors of PRP

results. Dependent variable was change in disability (PDI T1-T0). Independent variables were:

PDI baseline, PRP duration (weeks) and contact hours, gender, average pain at baseline and

interaction term of gender and duration (weeks). Dosage of PRP or any of the other variables did

not signifi cantly contribute to the regression model (results not presented).

Table 4 Differences between subgroups of patients, who completed the trial, according to duration of PRP, per protocol or shorter and extended

PRP per protocol or shorter PRP extended Mean Difference 95% CI P valuemean sd n mean sd n

PDI T0 37.7 12.3 77 33.0 14.0 37 4.7 0.42 to 9.79 0.07PDI T1 26.7 16.6 71 22.8 16.2 38 3.9 -2.67 to 10.49 0.24PDI T0-T1 10.4 10.4 68 9.4 16.0 35 1.1 -4.12 to 6.26 0.68

PDI: Pain Disability Index (scale 0-70), PRP: Pain Rehabilitation Program, T0: before treatment, T1: after treatment.

R1R2R3R4R5R6R7R8R9


68 | Chapter 4

Adverse events

No trial related adverse events were reported.

Discussion

The statistical analyses showed a point estimate difference between groups of 2.5 with CI from

-2.2 to 7.3, which means non-inferiority was inconclusive. Within both groups, short term ES

were moderate to large for outcome measures disability and quality of life. In a quarter of all

cases the duration of PRP could be reduced.

Because of extension of PRP in patients randomized to CAU-SF, this study was not able to detect

differences in dosage or effect. The results could imply that shortening of PRP duration may be

considered in some cases, without loss of short term results. A lower dosage may benefi t patients

and other stakeholders. However, 26% of all PRPs (36% within CAU-SF and 17% within CAU)

were extended to achieve a desired outcome. Extending (and shortening) of PRP was always

agreed upon by the patients and the PRP team, and was a result of a difference between baseline

needs and progress during PRP. At this point in time it is not possible to identify characteristics of

patients whose PRPs were extended or shortened, and thus to plan a correct dosage at the start.

This topic should be subject of further investigation. On the other hand, the results of the extended

subgroup were similar to the per protocol group, indicating that extension of treatment might

not be a solution to improve treatment outcome. Signifi cant differences between completers and

drop-outs may be hypothesized by less motivation of patients with disability compensation or

housewives to complete treatment. However, characteristics of dropouts have not been subject

of study. It is unknown whether this has systematically infl uenced the results of this study.

Because this trial is the fi rst of its kind, the results of this trial cannot be compared to other pain

rehabilitation trials with similar designs. Patient characteristics such as pain disability, pain intensity,

pain duration, gender distribution, appear similar to other PRP trials, both in the Netherlands

and internationally 5, and regular clinical secondary and tertiary rehabilitation programs in the

Netherlands. Dosage of PRP in this study was ‘midrange’ when compared to dosage reported

in other studies 14. The results of this study stress the relevance of further dosage studies in

different settings 5. One of the main challenges expressed in pain rehabilitation is the issue of

‘what works for whom’. We suggest adding another challenge: ‘how much works for whom?’

Within pain rehabilitation, this may apply to the multidisciplinary program as a whole, but also

to its components. This question may open a new line of research that may lead to major new

insights. Because dosage has been a methodological blind spot in PRP research, and dosage

variables were defi ned and interpreted differently across studies, results of previous trials may

have been biased by lack of control for dosage. If A is compared to B and A leads to superior



4

results, it is expected that A is the treatment of choice. However, theoretically, if the dose of A

was higher than B, the difference also might be explained by under dosage of B 15. Differences

in dosage could also explain differences between studies and could be reasons why some studies

were not able to detect signifi cant differences between interventions 14.

Weaknesses and strengths

Weakness of the study was that it had to be discontinued prematurely because too many PRPs

in de CAU-SF group were extended. This caused lack of contrast regarding dosage between the

CAU and CAU-SF groups, which may explain partly the inconclusiveness of the results. In addition

the small sample size caused wide CI, which also led to inconclusiveness.

Different scenario’s were considered and discussed with the team to shorten the program. CAU

was offered in programs of 8,12,16 and 20 weeks. In absence of evidence, it was decided to test

shortening of PRP with 1 step (4 weeks). This could be considered a methodological weakness

of the study, because the relative impact differed between programs (from 12 to 8 weeks: 33%

reduction, 12 to 16 week: 25% reduction, 20 to 16 weeks: 20% reduction). On the other hand,

adherence to CAU strengthens this study because it reduced the risk of allocation bias in patients

and clinicians.

Extension of treatment duration may be regarded as a limitation or even as adverse event or

program violation. However, during this study, we were obliged to adhere to the Dutch health

care regulations and the Medical Ethics Committee. As a consequence, we were not allowed

to deny health care that was deemed necessary for good patient care. Within the area of pain

rehabilitation there are no guidelines regarding dosage of PRPs. As described in the introduction,

PRPs aim to reduce disability and use of health care services. During PRP patients are coached to

self manage pain and disability. As far as we know there are no published validated measures

to assess these self management skills. Consequently, it is unclear at what point a patient will

be ready to self manage his pain and disability, and when PRP is no longer of added value.

Partly because of lack of evidence regarding dosage of PRP, choices of dosage of PRP are unclear

and arbitrary. This may result in differences in dosage between and within PRP teams. This is a

weakness of this study because, for some cases this resulted in shortening or extending of PRP

in both the CAU and CAU-SF groups. On the other hand, this is an important reason for future

research to focus on the rationale underlying dosage of PRP, including transparency of choices in

dosage of treatment. This should eventually lead to more rational dosage of PRP, including explicit

arguments on which dosage for individual patients is based. This could result in development of

validated measurements to determine dosage of PRP.

A strength of this study is the pragmatic, clinical design; including the struggle of clinical practice

regarding dosage of PRPs. This study shows results of shortening and extending PRPs. Based on

the regression analyses, which showed that dosage did not signifi cantly contribute to the model,

reducing PRP is possible and extending does not automatically lead to better results. As this

R1R2R3R4R5R6R7R8R9


70 | Chapter 4

was the fi rst trial within pain rehabilitation to study dosage, we suggest replication in different

settings, and that all future trials in this fi eld clearly describe dosage issues to enable future

(meta-)analyses of trial results.

In conclusion, reduction of dosage of PRP did not lead to non-inferior mean results. Because the

CI of the mean difference exceeded both the upper and the lower border of the non-inferiority

margin, the results of this trial are inconclusive.

Acknowledgements

The authors acknowledge the participation of the patients in this study, and the clinical and

support staff of the Pain Rehabilitation Department of the Centre for Rehabilitation of the UMCG.



4

References




4. Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain. Cochrane Database Syst Rev. 2014;9:CD000963.

5. Waterschoot FP, Dijkstra PU, Hollak N, de Vries HJ, Geertzen JH, Reneman MF. Dose or content? effectiveness of pain rehabilitation programs for patients with chronic low back pain: A systematic review. Pain. 2014;155(1):179-189.


7. Gronblad M, Hupli M, Wennerstrand P, et al. Intercorrelation and test-retest reliability of the pain disability index (PDI) and the oswestry disability questionnaire (ODQ) and their correlation with pain intensity in low back pain patients. Clin J Pain. 1993;9(3):189-195.

8. Soer R, Reneman MF, Vroomen PC, Stegeman P, Coppes MH. Responsiveness and minimal clinically important change of the pain disability index in patients with chronic back pain. Spine (Phila Pa 1976). 2012;37(8):711-715.

9. Herdman M, Gudex C, Lloyd A, et al. Development and preliminary testing of the new fi ve-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011;20(10):1727-1736.

10. Soer R, Reneman MF, Speijer BL, Coppes MH, Vroomen PC. Clinimetric properties of the EuroQol-5D in patients with chronic low back pain. Spine J. 2012;12(11):1035-1039.

11. Janssen MF, Pickard AS, Golicki D, et al. Measurement properties of the EQ-5D-5L compared to the EQ-5D-3L across eight patient groups: A multi-country study. Qual Life Res. 2013;22(7):1717-1727.

12. Borenstein M. Introduction to meta-analysis. Chichester, UK. Wiley; 2009:421.

13. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988:567.


15. Bendix T, Bendix A, Labriola M, Haestrup C, Ebbehoj N. Functional restoration versus outpatient physical training in chronic low back pain: A randomized comparative study. Spine (Phila Pa 1976). 2000;25(19):2494-2500.

FrrFrananankakaka PPP.C.CC. . WaWaW tetersrsrschchchooooo t,t,t, EEElsslselee innneee BeBeBennnnenenen MMMscscc, HeHeHenrnrnricicaaa R.RR SSSchchchipipphohohorsrsst tt PrPrPreueuupepeper r r MDMDMD,, PhPhPhD,D,D PiiPietetetererer UUU. DiDiDijkjkkststs raaa PPPhDhDhD, JaJaJann H.H.H.B.BB. GGGeeeeeertrtzezez nnn MDMDMD, PhPhPhD,D,D, MMMiccchihihielele FFF. ReReReneneemamaann PhPhPhDDD

SuSuubmbmbmititittetted dd fofofor rr pupupublbbliciicatatatioioionnn

R1R2R3R4R5R6R7R8R9


74 | Chapter 5

Abstract

There is a large variety in prescribed dosages of pain rehabilitation programs (PRP) but evidence

regarding the optimum dosage is unknown. The aim of this study was to explore perspectives

of patients and rehabilitation professionals regarding dosages (e.g. total duration, number of

contact hours) of PRP in three different rehabilitation centers in the Netherlands. A study using

an explorative qualitative research design was performed with thematic analysis. Individual semi-

structured interviews were conducted with patients undergoing PRP, and focus groups were

formed with rehabilitation professionals involved in PRP. In total, 12 patient interviews and three

focus groups with 17 rehabilitation professionals were analyzed. All patients were satisfi ed with

offered dosage. Several factors were considered important in relation to the choices in dosage of

PRP. All factors were categorized as patient-related factors, treatment-related factors and external

factors. Although dosage of PRP differed among the participating rehabilitation centers, patients

and rehabilitation professionals mentioned similar factors. In absence of evidence, differences in

choices of PRP dosage appear mainly based on historical grounds and clinical expertise.

Keywords

Pain rehabilitation programs, dosage, qualitative analysis, chronic pain.


Patient and professionals’ perspectives on dosage of PRP | 75

5

Introduction

Multidisciplinary Pain Rehabilitation Programs (PRPs) are recommended to treat patients with

chronic musculoskeletal pain (CMP) 1,2. These PRPs, based on the biopsychosocial model, aim to

decrease disability and optimize participation of patients with CMP. A systematic review showed

that PRPs have a moderate, but consistent, positive effect on disability and pain, compared to

usual care or physical treatment programs for patients with chronic low back pain 3. PRPs result in

better self-management of pain and disability, and a reduction in healthcare utilization in treated

patients was found, which may contribute to a decrease of direct and indirect costs over the long

term 4.

Although effective, PRPs are relatively expensive. The multidisciplinary characteristics, as well

as the high number of contact hours and total duration of PRP, provide relatively high direct

costs and travel expenses for the patients. An aspect of PRP is dosage, which includes the total

duration, the total number of contact hours, and intensity of treatment (number of contact hours

per week). Differences in the dosage of PRP may lead to differences in direct and indirect costs.

Choices for dosage and how dosage is established for an individual patient is unknown. A recent

systematic review showed that dosage of PRP has never been studied as a primary aim and the

optimum dosage of PRPs is currently unknown 5. The studies included in that review differed in

terms of dosage (total duration, from weeks to months, and regarding contact hours, from fewer

than 10 to over 100 hours) and effect.

Better understanding of dose variables could lead to better and more effi cient patient care, which

will benefi t patients, rehabilitation facilities, insurers and employers. To acquire insight into factors

related to dosage of multidisciplinary PRP, this study aimed at exploring perspectives of patients

who underwent PRP, as well as of the rehabilitation professionals working in PRPs. In addition,

this study aimed at examining the argumentations/reasons and underlying choices in dosage of

PRPs in three rehabilitation centers in the Netherlands.

Methods

A study using an exploratory qualitative design with thematic analyses was used to gain in-depth

information about the perspectives of patients and rehabilitation professionals involved in PRPs

regarding the doses of these programs. The study was performed in three rehabilitation centers in

the Netherlands. These centers were selected because of their differences in dosage, while patient

characteristics were similar 6. Individual semi-structured interviews with patients who underwent

multidisciplinary PRP, as well as focus groups with professionals working in multidisciplinary PRP,

R1R2R3R4R5R6R7R8R9


76 | Chapter 5

were used. The medical ethics committee of the University Medical Center of Groningen (UMCG)

granted a waiver for this study. All participants signed informed consent for participation, digital

recording and the use of data.

Participants

A purposeful sample of patients was recruited from three rehabilitation centers in the Netherlands:

Adelante in Hoensbroek (RC1), located in the south; Roessingh in Enschede (RC2), located in the

east and the Center of Rehabilitation of the UMCG (RC3), located in the north. We aimed at

including patients with varying characteristics, who completed a multidisciplinary PRP: working/

not working, differences in dosage of PRPs, males/females, and a variety of ages. PRP should

have been completed for at least two weeks. The fi rst author contacted the recruited patients via

telephone to check their willingness to participate in the study and to plan the interview. They

received written information about the purpose of the interview.

Rehabilitation professionals from the three rehabilitation centers were included for participation

in the focus groups, provided that they were working within the area of multidisciplinary PRP for

a minimum of two years, for at least 0.5 fte. We aimed to include rehabilitation professionals

from different disciplines (i.e. occupational therapist/ physiotherapist/ psychologist/ physiatrist).

These rehabilitation professionals should be involved in determining the individual dosages of the

program for the patient.

Data collection

Data were collected between May and September 2014. A total sample of 15 patients was

planned for the interviews, fi ve patients from each rehabilitation center, with the expectation

of reaching data saturation. A semi-structured interview scheme was constructed [Appendix]

for patients using open questions enabling refl ection about experiences and perceptions on

dosage of treatment received. The interview started with questions about the total duration

and intensity of PRP, followed by questions regarding personal experiences and perceptions of

treatment dosage, such as: “what did you like and dislike about the duration and intensity of the

program” and “which factors, in your opinion, infl uence dosage of PRP?”. Also patients’ opinion

regarding the relationship of dosage to return to work and personal and general costs were

asked. Face-to-face interviews were performed at the rehabilitation center where the patient was

treated. The interviews were scheduled for 60 minutes and were audio recorded. Prior to this

study, an interview was pilot tested. Patient interviews were planned prior to the focus group at

each center.

For the focus groups meetings, a minimum of fi ve rehabilitation professionals of different

disciplines per center were invited in order to achieve variation in perspectives on PRP dosage per



5

rehabilitation center. Meetings were planned for 90 to 120 minutes and were audio recorded. An

interview scheme was constructed [Appendix], aiming to gain insight into the line of reasoning

underlying dosage choices, as well as rehabilitation professionals’ perspectives on optimum

dosage of PRPs.

Measures of validity

The fi rst author (FW), who has over 10 years of experience as an occupational therapist in pain

rehabilitation, conducted all patient interviews and also led the focus group meetings in the

Adelante and Roessingh rehabilitation centers. Because she was closely acquainted with the

rehabilitation professionals at UMCG, a psychologist with experience in chronic pain treatment

but who was not involved in PRP of the UMCG led this focus group. The second author (EB)

assisted in all focus groups.

Data analysis

Interviews and focus groups were audio recorded and transcribed verbatim by the fi rst and

second authors (FW and EB). Transcripts were imported into ATLAS.ti, version 5.2, to analyze

data. Thematic analysis was considered as the appropriate approach for identifying, analyzing

and reporting factors related to dosage of PRP 7. Because of the lack of evidence regarding

dosage of rehabilitation programs, especially PRP aimed at behavior change, it was not possible

to provide a detailed theoretical framework. Initially, the fi rst and second authors (FW and EB)

conducted the inductive coding of the interviews and another independent researcher validated

the codes. All codes and quotations were re-analyzed, merged, and renamed.

Results

Participants

In total, 13 patients were interviewed and 12 interviews were analyzed. Initially, fi ve patients

per center were planned, two dropped out because of logistics problems and one was excluded

because the audio was impossible to transcribe. Seventeen rehabilitation professionals participated

in the focus groups (RC1: 5; RC2: 5; RC3: 7). Different disciplines participated: occupational

therapists (4), physiotherapists (3), movement teacher (1), psychologists (4), social workers (2)

and physiatrists (3). The characteristics of participants are described in Table 1. Data saturation

was reached for patient interviews and focus groups for rehabilitation professionals.

R1R2R3R4R5R6R7R8R9


78 | Chapter 5

Table 1 Characteristics of participants and PRP dosage

RC1 RC2 RC3(n=4) (n=4) (n=4)

PatientsAge mean (SD) 42 (16) 37 (11) 53 (7)

Female % 75 100 100Duration of pain n

1 to 2 years 0 1 2 2 to 5 years 2 0 0 > 5 years 2 3 2

Initial PDI score mean (SD) 27 43.3 (14.2) 31.5 (16.4)Missing n 3 0 0

Rehabilitation professionals(n=5) (n=5) (n=7)

Age mean (SD) 41 (6) 43 (9) 39 (6)Female % 60 80 71Years of experience in PRP mean (SD)

7.6 (4.7) 9.0 (5.6) 7.4 (5.5)

PRP Range in total duration (wks)

Assessment 3 1 to 2 1 to 2Treatment 12 3 to 36 8 to 20

Range in contact hoursAssessment 30 2 3Treatment 120 10 to 150 30 to 70

PDI: Pain Disability index, PRP: Pain Rehabilitation Program; RC1: Adelante in Hoensbroek; RC2: Roessingh in Enschede, RC3: Center for Rehabilitation of the UMCG

Factors related to dosage

Different codes related to dosage of PRP were derived from the interviews and focus groups.

Codes were categorized as patient-related factors, treatment-related factors or external factors. In

addition, the codes were arranged as “shared” (indicating that they were mentioned by patients

and rehabilitation professionals), “patients” (those that were mentioned by patients only), and

“rehabilitation professionals” (those mentioned by rehabilitation professionals only) (table 2).



5

Tab

le 2

Ove

rvie

w o

f co

des

per

cate

gory

Shar

ed f

acto

rsFa

cto

rs f

rom

inte

rvie

ws

wit

h

pat

ien

tsFa

cto

rs f

rom

fo

cus

gro

up

s w

ith

re

hab

ilita

tio

n p

rofe

ssio

nal

s

Pati

ent

rela

ted

fac

tors

Cas

e co

mpl

exity

Ass

ertiv

enes

s-

Gen

eral

sta

tus

Rela

pse

Inse

curit

ySe

lf-kn

owle

dge

Mot

ivat

ion

and

abili

ty t

o ch

ange

beh

avio

rEx

pect

anci

es

Focu

s on

you

rsel

fA

ccep

tanc

e of

pai

nSe

lf m

anag

emen

tA

pply

less

ons

lear

ned

into

pra

ctic

eTe

am &

tre

atm

ent

rela

ted

fac

tors

Wai

ting

time

Hav

ing

time

and

oppo

rtun

ity t

o ex

plor

eEh

ealth

Cla

rity

abou

t do

sage

and

tim

e pl

anSa

tura

tion

Lack

of

evid

ence

Indi

vidu

ally

tai

lorin

gEx

pert

ise

of t

he t

eam

Pred

ictio

n of

dos

age

of P

RPSh

ared

dec

isio

n m

akin

gTi

min

g to

fi ni

sh P

RPC

onta

ct w

ith r

ehab

ilita

tion

prof

essi

onal

sIn

terd

isci

plin

ary

func

tioni

ngC

onte

nt o

f tr

eatm

ent

Trea

tmen

t go

als

Form

at o

f tr

eatm

ent

Exte

rnal

fac

tors

Dire

ct c

osts

Supp

ort

from

env

ironm

ent

Trav

ellin

g tim

eIn

dire

ct c

osts

Form

at o

f tr

eatm

ent

Inve

stm

ent

for

the

futu

reIn

jury

com

pens

atio

n Pe

rson

al a

nd w

ork

fact

ors

Test

res

ults

or

othe

r tr

eatm

ents

R1R2R3R4R5R6R7R8R9


80 | Chapter 5

Patient-related factors:

Shared factors:

Both patients and rehabilitation professionals mentioned case complexity of a patient as an

important factor related to dosage of PRP as well as an important issue to take into account for

indication of dosage of PRP. Both patients and rehabilitation professionals described characteristics

that affect dosage of PRP, such as the general status (the level of physical condition or the

degree of experienced disability), as well as the degree of insecurity, self-understanding,

motivation and ability to change behavior, in addition to the expectancies patients have

regarding treatment and treatment results. (P3): “It depends on what is going on with someone.

Is he traumatized? Does he want to talk about it? Is he able to manage his problems? Does he

have some self-understanding; does he know what his problem is?”

Both patients and rehabilitation professionals mentioned that PRP requires focusing on

yourself. Patients and rehabilitation professionals involved with both high and low PRP dosages

mentioned this focus as being related to dosage. A patient (P7) who underwent semi-inpatient

PRP described: “The inpatient format of treatment was a good experience. It was quite intense,

but actually very good because you really have to focus on yourself. If I would have been at

home, I would have been distracted by all the things you have to do at home”. Focusing on

oneself and on the treatment was also mentioned by a patient (P9) as motivation for preferring

twice a week PRP instead of once a week: “Yes, two times a week, than you are really engaged

with the treatment. Treatment only once a week…. well, then a whole week is in between…”.

Rehabilitation professionals mentioned that being able or motivated to focus on oneself results

in being motivated to follow the prescribed dosage of PRP. In some cases, for example (T13) “if

a patient is overloaded”, being able to focus on oneself is a motivation for the choice of dosage

and the format of PRP, such as inpatient PRP.

Acceptance of pain is related to the content of PRP because PRP is aimed at improving function,

not reducing pain. Participants mentioned that PRP contributed to the acceptance of pain and

that patients need time to go through this process of acceptance (P4): “I think the minimum

duration of PRP should be 12 weeks. Especially because you experience a process of change in

which you need time… time to make a start, because after those 12 weeks you still need to go

on for yourself”.

If a patient is able to self-manage his pain and disabilities, participants suggest needing a lower

dosage of PRP compared to a patient that is not able to. A rehabilitation professional in a focus

group (T16) described: “Sometimes the minimum dosage of PRP is just one hour of intake. There

are patients who just need some explanation to self manage their lives again. Being able to self

manage pain is related to patient characteristics and behavior, the way people are, and whether



5

they are able to cope with their pain and disabilities themselves”. Patients also mentioned that

if they are able to learn the principles of self-management during the process of PRP, they will

ultimately need a lower dosage.

To apply lessons learned into practice, patients need time. This is a motivation of patients and

rehabilitation professionals for dose variables such as total duration of PRP. (P9) “It is a process

which requires time, 6 or 8 weeks… well, that’s quite short…. You need to be aware of things,

you have to practice, adjust and adopt”. Needing time to apply lessons learned was additionally

mentioned with regard to intensity (contact hours per week) and the reduction of intensity

of contact hours during PRP. (T13): “At the end of the program, reducing the intensity of the

program (fewer days or fewer contact hours per week), could support the process of putting the

lessons learned into practice”.

Factors from interviews with patients:

Assertiveness was mentioned by patients with regard to the ability to stand up for themselves.

This code is part of the goals and content of PRP and therefore patients mentioned this in relation

to needing a lower dose when patients are already able to stand up for themselves. However, the

code has also been reported in terms of “shared decision making” regarding dosage of PRP.

(P8) “I had to learn to set boundaries. My day was fully planned and I used to do whatever I was

told, which resulted in being exhausted at the end of the day”. Some patients reported that they

needed support during times of relapse. (P2): “Everything was going well; I also felt it that way.

However, I did not have the feeling that I was ready. I was afraid that I would relapse and that

was why I needed some sessions”.

Factors from focus groups with rehabilitation professionals:

No other patient-related factors were reported by rehabilitation professionals only.

Treatment-related factors:

Shared factors:

Waiting time before and during treatment can infl uence, in a negative or positive way, the

general status of a patient at the start of treatment, therefore having an infl uence on the dose

of PRP. One patient (P1) described how she was able to turn the negative result of the waiting

time into a positive action: ”I had to wait for a long time before start of PRP. For me personally,

it has led to an increase of pain and I became depressed. Consequently, I realized I had to do

something and I started doing things on my own”. Most patients knew the total duration of PRP

beforehand. Patients were satisfi ed with this clarity about dosage and time plan for PRP: (P4)

“I compare it with knowing when to go on holidays: the last week at work you are really into

holidays. With therapy, it is the same – you have a set goal to work on in time”. Nevertheless,

R1R2R3R4R5R6R7R8R9


82 | Chapter 5

besides structured dosages providing this clarity, participants emphasized that the possibility

of individually tailoring dosage to their needs is important. The individually tailored dosage

is also related to shared decision-making. Patients and rehabilitation professionals attached

value to shared decision-making in terms of PRP dosage. Good contact between patient and

rehabilitation professionals supported therapy and therefore could reduce dosage: (P11): “You

need to have a connection with your rehabilitation professionals; otherwise you will not make

any progress”. However, too many changes in rehabilitation professionals during therapy can

impede the connection, and therefore the process, because: “you have to tell your story again

and again”. Related to this code, patients and rehabilitation professionals mentioned that good

interdisciplinary functioning of the team supported effectiveness and therefore had a positive

infl uence on PRP dosage. (P9): “I experienced that everyone knew who I was and what I needed.

Everyone had read all reports and you did not need to tell the same stories”. Both patients and

rehabilitation professionals made it clear that dosage of PRP was strongly related to the content

of treatment, treatment goals and the format of treatment (e.g. inpatient versus outpatient).


Most patients had the experience of having time and opportunity during PRP to explore

themselves and reach their goals. However, some patients also experienced this as a problem when

the frequency was reduced from fi ve days to two days per week. “Content became more intensive,

everything was put together in two days and everything had to go faster…I had to hurry….”.

Some patients denoted that towards the end of treatment, saturation of treatment content

occurred. Therefore, they suggested that retrospectively, the dosage could have been reduced.

(P2): “At a certain moment, you are done with it”. Patients stated that they trust the expertise

of the team, both with regard to the content of treatment as a prediction of dosage and when

decisions are made regarding the need to prolong or reduce PRP dosage. (P1): “I thought, well if

that is the success of the therapy, then I am going to do it that way”. Rehabilitation professionals

felt that they had to be experts in the fi eld and should offer a high PRP quality. They related this

quality and expertise to content and dosage of PRP.


Rehabilitation professionals suggested that different forms of ehealth could infl uence the

dosage of PRP in the near future. “Patients can receive home assignments by mail after intake

and before starting. Also aftercare can be supported by ehealth”. The focus groups revealed that

there was a lack of evidence for the delivered doses of PRP. (T15) “These programs have been

developed and adapted over the years based on skills acquired by the team, and developments

such as group based treatments. This provides certain components in the program. This is what

we have; we did not question ourselves if it is necessary for patients”. Additionally, with that lack

of evidence, it is hard to predict dosage of PRP and to know when the timing is right to fi nish



5

PRP: “It is a diffi cult decision, at what point a patient is ready, and to reach consensus about this

within the team and also with the patient…. “

External factors:

Shared factors:

Opinions regarding fi nancial factors were of interest to the interviewer. The majority of participants

stated that fi nancial factors should not infl uence dosage of PRP. Patients could imagine that

direct costs that patients have to pay themselves, like travel costs, could be of concern for some

patients, although they did not experience that for themselves. (P11): It is regrettable if your own

development depends on a few Euros extra per month”. Direct costs of therapy and indirect

costs, for example those related to returning to work, should not infl uence treatment because

patients saw the costs of therapy as an investment in the future: (P9): Well, I think, on the

longer term, if therapy works, you save money. It also applies for work: if my boss did not allow

me time for this therapy, I might not yet be at work for 100%”. Personal and work factors

could have a positive or negative infl uence on the opportunity to receive therapy, and therefore

could infl uence the satisfaction or choice for dosage of PRP. Some patients had to adjust their

activities at home to their therapy: (P11): “On therapy days, I did not have the energy to do things

at home. I was happy I made the choice to delegate the household tasks. The days at home I had

time to do things with my children and husband”. A rehabilitation professional mentioned that

sometimes, dosage of therapy can be adjusted to work or personal situations: “If patients are

not able to organize the required dosage of two times a week and they have strong reasons, we

sometimes decide to reduce frequency of treatment to once a week”.


Patients perceived support from environment (e.g. partner, colleagues, employer) as a positive

factor because it helped with completing the program well and spending the time needed for the

program. Lack of support could have a negative infl uence on the effects of the program because

it could interfere with the process of therapy and lead to extension of PRP.


In the focus groups, rehabilitation professionals mentioned that travel time could infl uence

the choice of the offered format of treatment, therefore infl uencing the dosage of PRP. “This

center is a center of expertise regarding pain rehabilitation, so patients come from across the

province, which makes outpatient rehabilitation not appropriate”. If the patient received injury

compensation, it sometimes can impede progression in treatment. Therefore, rehabilitation

professionals stated that in some cases, this could be a reason for not starting PRP or for stopping

it prematurely. Test results or other treatments can also infl uence the dosage of PRP: “More

often you see a sort of intermediate treatment if patients need other treatments, like EMDR

R1R2R3R4R5R6R7R8R9


84 | Chapter 5

….. Or if test results are not available; waiting for these results causes delay” (EMDR is the

abbreviation for Eye Movement Desensitization and Reprocessing; trauma treatment).

Differences and similarities

Interviews and focus groups were performed in three rehabilitation centers across the Netherlands.

All centers offered multidisciplinary PRP for patients with CMP. Although dosages differed per

center, factors were similar for all centers. All patients, in general, mentioned being satisfi ed

with the received dosage, although some had suggestions for improvements related to dosage of

PRP. Rehabilitation professionals reported that they supported the choices for current dosages of

PRP; however they also had suggestions for improvements and implications for further research

in order to eliminate the lack of evidence for an optimal dosage of PRP.

Discussion

As a contribution to gather knowledge about factors related to dosage of PRP, this study provides

insight into perspectives of patients and rehabilitation professionals regarding dosage of PRP.

Several general factors were considered relevant. Overall, all codes could be summarized into

patient, treatment and external factors. Interestingly, despite differences in offered and received

dosages of PRP, these factors were similar across centers with different PRP dosages. For example,

in general, patients need clarity about dosages at the start of PRP. However, during PRP, they

would like to have the opportunity to adapt the dosage to their individual needs, in consultation

with their rehabilitation professionals. Patients and rehabilitation professionals mentioned that

different personal characteristics, such as motivation and ability to change, general status of

a patient, acceptance, and self-understanding, are related to dosage of PRP. However, similar

“ability to change” can lead to 12 weeks, divided into treatment of six weeks, fi ve times a week,

eight hours a day, and six weeks, three times a week, eight hours a day of PRP in RC1, and 12

weeks, twice a week, 2,5 hours a day of PRP in RC3. Remarkably, no explicit reasons for these

differences in choices of dosage were expressed during the focus groups.

Treatment outcome is caused by specifi c and aspecifi c factors of treatment. The content of PRP

is a specifi c factor. Outcome, defi ned as “being satisfi ed” with treatment, is an example of an

aspecifi c factor infl uencing the effect of treatment. The factors in this study regarding dosage

of treatment seemed to be linked both to specifi c and aspecifi c factors. Ehealth and evidence

regarding optimum dosage are specifi c factors which can contribute to better prediction of

dosage and more effi cient PRP. Content and format of treatment, and external factors (e.g. direct

and indirect costs, investment in future, personal and work factors, support, test results) are

specifi c factors infl uencing choices of, and satisfaction with dosage. These specifi c factors could



5

explain part of the differences in choices of the offered PRP dosage. However, the results of this

study showed that also other, more aspecifi c factors seemed to be relevant although not specifi c

related to dosage.

Some of these latter factors are similar to the common factors analyzed for explaining equivalence

of various treatment approaches 8,9. In a systematic review fi ve categories of common factors

were analyzed in psychotherapy research: patient characteristics, therapist qualities, change

processes, treatment structure and therapeutic relationships 8. Patient characteristics, such as

expectations of patients with chronic pain, are related to outcomes of pain treatment 10-12.

Rosenzweig 9 already in 1934 described the effect of therapy characteristics with the introduction

of the concept of common factors. In physical therapy care, a systematic review analyzing

patient satisfaction concluded that characteristics of the therapist and the process of care were

determinants of satisfaction of musculoskeletal physical therapy care 13. However, the association

between satisfaction and treatment outcome was inconsistent. The category of change processes

consists of, for example, the opportunity for catharsis, practice of new behaviors in a safe

environment, provision of rationale and foster insight 8. These common factors are similar to

the factors identifi ed in this study: having time and opportunity to explore, applying lessons

learned into practice, acceptance of pain, and self-management. Hush et al. 13 also described the

relationship between patient satisfaction and the process and organization of care. Timely and

effi cient treatment, adequate treatment frequency and follow up, patient involvement in decision

making and individualized care were variables of the process of care, contributing to patient

satisfaction. Variables contributing to patient satisfaction regarding the quality and effi ciency

of the organization of care were: good access to services, location, parking and approachable

support staff 13. Communication, as part of the treatment structure 8, is concluded to be an

important factor in healthcare 14. This is in line with the factors “clarity”, “shared decision-

making” and “contact with rehabilitation professionals” mentioned by patients and professionals

in this study. Communication based on patient-centered care strategies, is also associated with a

stronger therapeutic relationship 15. In addition, studies have analyzed therapeutic relationships as

important common factors that could be related to the effects of different healthcare treatments 16-18. All common factors were analyzed in different contexts of treatments and in the associations

between different variables of treatment. Remarkably, however, in this study with similar contexts

and aims of PRP, similar common factors led to different dosages of PRP. Therefore, it is a challenge

to analyze the effect of the different components of PRP on outcome.

As a result of the lack of evidence regarding dosage of PRP, as mentioned in the focus groups,

all rehabilitation centers made different choices regarding dosage. These choices were based

on different motives related to content, dosage, and composition at the time of developing

the multidisciplinary PRP. Over time, dosage was adapted because of changes in society, costs,

R1R2R3R4R5R6R7R8R9


86 | Chapter 5

experiences, and new insights related to content or format of therapy. At present, no evidence

can support the choices of dosage of PRP and therefore, rehabilitation professionals are not able

to rationally choose optimum dosage of PRP.

Strengths and limitations

Analyzing factors associated with dosage using a qualitative method from the perspective of

patients and rehabilitation professionals is considered a strength of this study because it is the

fi rst study providing an overview of factors related to dosage from patient and professionals’

perspectives. Taking into account three different rehabilitation centers with different dosages of

PRP and similar content across the Netherlands is a strength because this provides insight into the

differences and similarities regarding perspectives of dosage, resulting in similar factors regardless

of variety in the offered dosage. Although this study was performed in three different centers, it

focused solely on centers in the Netherlands, limiting generalization to PRPs performed in centers

in other countries. In addition, the qualitative nature of this study resulted in perspectives of

patients and rehabilitation professionals and cannot simply be generalized to all other groups

of patients and rehabilitation professionals involved in PRPs. For example, we were not able to

include more male patient participants. Although, the majority of patients in PRP are female,

more male patient participants would have improved comparability to clinical practice.

Recommendations

This study is an initial exploration of the perspectives of patients and rehabilitation professionals

regarding PRP dosage. It resulted in an overview of factors that can be used for future research

and clinical practice regarding PRP dosage. The overview can contribute to further analysis of

these factors and their relationship to dosage. Also, it can improve the awareness of these factors

related to dosage of PRP in clinical practice. The results indicate that factors other than content

and dosage of PRP were perceived as being important factors by patients and rehabilitation

professionals. Clinical practice can be improved by taking these other factors into account as

contributors to the overall success of PRP.

Conclusion

This study shows that, although dosage of PRP differed, patients and rehabilitation professionals

generally mentioned the same factors related to characteristics of patients and treatment, as well

as similar external factors. In absence of evidence, differences in choices of PRP dosage appear

mainly based on historical grounds and clinical expertise. Therefore, research is needed to guide

choices of optimum PRP dosage.



5

Acknowledgements

The authors would like to thank all patients and rehabilitation professionals who participated

in the interviews and focus groups at all three rehabilitation centers. They also extend their

appreciation to Ernst Schrier for leading the focus group and Annemieke de Jong and Suzanne

Broekema for their support with data analysis.

Confl ict of interest

We certify that no party having a direct interest in the results of the research supporting this

article has or will confer a benefi t on us or on any organization with which we are associated.

R1R2R3R4R5R6R7R8R9


88 | Chapter 5

References

1. Turk DC. Clinical effectiveness and cost-effectiveness of treatments for patients with chronic pain. Clin J Pain. 2002;18(6):355-365.

2. Airaksinen O, Brox JI, Cedraschi C, et al. Chapter 4. European guidelines for the management of chronic nonspecifi c low back pain. Eur Spine J. 2006;15 Suppl 2:S192-300.

3. Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ. 2015;350:h444.

4. Lambeek LC, van Tulder MW, Swinkels IC, Koppes LL, Anema JR, van Mechelen W. The trend in total cost of back pain in the Netherlands in the period 2002 to 2007. Spine (Phila Pa 1976). 2011;36(13):1050-1058.


6. Preuper HR, Boonstra AM, Wever D, et al. Differences in the relationship between psychosocial distress and self-reported disability in patients with chronic low back pain in six pain rehabilitation centers in the netherlands. Spine (Phila Pa 1976). 2011;36(12):969-976.

7. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3(2):77-101.

8. Grencavage LM, Norcross JC. Where are the commonalities among the therapeutic common factors? Professional Psychology: Research and Practice. 1990;21(5):372-378.

9. Rosenzweig S. Some implicit common factors in diverse methods of psychotherapy. Journal of Psychotherapy Integration. 2002;12(1):5-9.

10. Kalauokalani D, Cherkin DC, Sherman KJ, Koepsell TD, Deyo RA. Lessons from a trial of acupuncture and massage for low back pain: Patient expectations and treatment effects. Spine (Phila Pa 1976). 2001;26(13):1418-1424.

11. Bhana N, Thompson L, Alchin J, Thompson B. Patient expectations for chronic pain management. J Prim Health Care. 2015;7(2):130-136.

12. Burns JW, Nielson WR, Jensen MP, Heapy A, Czlapinski R, Kerns RD. Specifi c and general therapeutic mechanisms in cognitive behavioral treatment of chronic pain. J Consult Clin Psychol. 2015;83(1):1-11.

13. Hush JM, Cameron K, Mackey M. Patient satisfaction with musculoskeletal physical therapy care: A systematic review. Phys Ther. 2011;91(1):25-36.

14. Roberts L, Bucksey SJ. Communicating with patients: What happens in practice? Phys Ther. 2007;87(5):586-594.

15. Pinto RZ, Ferreira ML, Oliveira VC, et al. Patient-centred communication is associated with positive therapeutic alliance: A systematic review. J Physiother. 2012;58(2):77-87.

16. Burns JW, Glenn B, Lofl and K, Bruehl S, Harden RN. Stages of change in readiness to adopt a self-management approach to chronic pain: The moderating role of early-treatment stage progression in predicting outcome. Pain. 2005;115(3):322-331.

17. Ferreira PH, Ferreira ML, Maher CG, Refshauge KM, Latimer J, Adams RD. The therapeutic alliance between clinicians and patients predicts outcome in chronic low back pain. Phys Ther. 2013;93(4):470-478.



5

18. Degnan A, Seymour-Hyde A, Harris A, Berry K. The role of therapist attachment in alliance and outcome: A systematic literature review. Clin Psychol Psychother. 2014.

FrFrF ananankakak PP.CCC. WaWaWatetetersrsrschchchoooooot,t,t, EEElslsselele innne e BeBeBennnnnnenenen MMMsccsc,,, LuLuLucacacas ss H.HH V.VV vvananan dddereer WWWououo dedede PPPhDhDhD,HeHeHenrnn iccaaa R.R.R. SSSchchchipipiphohohorsrsrst t PrPrreueeupepep r rr MDMDMD,, PhPhPhD,D,D, MMMiccchihhielele FFF. ReeRenenemamannn PhPPhD.DD

AcAcccececeptptptededed IIntntnt JJ RRRehehehabababiliil RRReseses

R1R2R3R4R5R6R7R8R9


92 | Chapter 6

Abstract

Assessment of case complexity of patients with chronic non-specifi c musculoskeletal pain (CMP)

is currently clinician based, not transparent and with low reliability. The objective of this study was

to explore case complexity and to initiate development of a Case Complexity Index (CCI).

A three-round Delphi study among clinicians involved in multidisciplinary Pain Rehabilitation

Programs was performed to identify important factors which are assumed to infl uence functioning

in patients with CMP. The 10 most important factors were used to initiate development of a CCI

with the mean ratings of importance per factor as weights. The feasibility of the CCI was tested

in a pilot study on 16 patients with CMP.

In the fi rst round, 166 factors were identifi ed; in the second round, the 10 most important factors

were selected; in the third round, relative weights of each factor were calculated, ranging from

1.75 (features of complaints) to 3.56 (psychiatric disorders) on a scale of 0 (no weight) to 4 (very

heavy weight). The assessments for the factors were mainly based on clinical examination and

reasoning. Clinicians could rate all patients using the CCI, which establish feasibility of the CCI.

Ten, mainly psychosocial factors were identifi ed, which were assumed to be most important for

assessment of case complexity of a patient with CMP. With these factors a CCI was created for

which feasibility was established. This CCI is transparent, easy to use and might provide a basis

for further developments of a structured assessment of case complexity, which may have scientifi c

and clinical relevance.

Keywords

Chronic pain, chronic non-specifi c musculoskeletal pain, multidisciplinary pain rehabilitation

program, case complexity, complexity index, Delphi technique


Case complexity of patients with CMP | 93

6

Introduction

Development of chronic non-specifi c musculoskeletal pain (CMP) is often explained with the

biopsychosocial model, because interacting factors may be involved infl uencing CMP and its

consequences 1,2. The number of factors and their relative involvement creates a highly variable

case complexity of patients with CMP.

Due to the multifactorial nature of the problem, a multidisciplinary approach is needed to treat

patients with CMP. Multidisciplinary Pain Rehabilitation Programs (PRP) are effective in improving

pain and daily functioning in patients with CMP 3, but improvement is modest. PRPs are highly

variable and optimal content and dosage are unknown 4. To analyze the clinical situation of

patients eligible for PRP, to set goals and to estimate dosage of PRP, clinicians assess the case

complexity of patients. This study defi ned case complexity as the extent to which combinations

of factors infl uence functioning of patients with CMP. It comprises both a number of factors

and their relative weight. Theoretically, a patient presenting a large number of factors that each

weigh heavily, is considered a more complex case than a patient with fewer factors that weigh

lighter. Complex cases might need different treatment and different dosage of PRP than low

complex cases. A patient presenting only one heavy weight factor could need high dosage of

mono-disciplinary therapy, but will not be considered as complex case for this study, because

there is only one factor infl uencing functioning. This patient is not eligible for PRP. Although

never investigated, for the evaluation of PRPs, case complexity could be considered as potential

moderator, because besides infl uencing content of treatment it might also infl uence dosage.

However, we are presently unable to reliably and validly assess case complexity, and therefore

unable to test relationships between case complexity and dosage of PRP.

Case complexity of patients admitted to PRPs, is currently assessed ‘intuitively’ by clinicians, partly

based on validated questionnaires. However, psychometric properties of the fi nal complexity

assessment are unknown. A standardized, transparent and psychometrically sound method

assessing case complexity specifi cally for patients with CMP is unavailable. In the Netherlands,

the Workgroup Pain Rehabilitation Netherlands (WPN) developed a four level classifi cation system

for case complexity of patients with CMP 5 (textbox 1). Patients with WPN level 3 or 4 are eligible

for PRP. While this classifi cation is considered as gold standard for the complexity assessment by

rehabilitation specialists in the Netherlands, its criteria are not described, lack transparency and

objectivity, and has moderate interrater reliability 6. Moreover, differentiation is low because levels

are very wide, and PRP content and dosage cannot be derived from this classifi cation. Thus, a

standardized, transparent and psychometrically sound method to assess case complexity will help

improve PRP scientifi cally and clinically.

R1R2R3R4R5R6R7R8R9


94 | Chapter 6

Textbox 1 Defi nitions of WPN levels (author’s translation)

WPN level Criteria1 Patients with pain and (imminent) limitations in activities without infl uence on (work)

participation, no or minimal psychosocial problems.

2 Patients with pain and limitations in activities and (work) participation, no or minimal psychosocial problems.

3 Patients with pain and limitations in activities and (work) participation, psychosocial problems that have moderate to severe infl uence on the (self-reported) level of functioning.

4 Patients with pain and limitations in activities and (work) participation, psychosocial problems that have very severe infl uence on the (self-reported) level of functioning.

The objectives of this study were to:

1. Explore the most important factors infl uencing case complexity of patients with CMP eligible

for PRP.

2. Initiate development of a Case Complexity Index (CCI) based on factors, deemed most

relevant by clinicians.

3. Explore means of assessment used to determine the relevance of CCI individual factors.

4. Test the CCI feasibility.

Materials and Methods

A 3-round online Delphi study was conducted among clinicians working in PRPs in the Netherlands.

Participants

Rehabilitation specialists were selected through purposive sampling and invited by email. The

rehabilitation specialists were asked to invite physiotherapists, psychologists and occupational

therapists from their rehabilitation team to participate (snowball sampling). Rehabilitation

specialists, physiotherapists, psychologists and occupational therapists who were for more than

two years substantially (0.5 FTE or more) involved in a PRP for patients with CMP were considered

experts for this Delphi study. The experts worked in rehabilitation centres spatially dispersed across

the country. For the sample size in a Delphi study there are no strict rules, but a minimum sample

size of 10-15 for a homogenous group of experts has been suggested 7. Because the sample

consisted of four different disciplines, a minimal sample size of 10 in each discipline was chosen.

First round

Objective: to determine all factors that are assumed to infl uence the case complexity of patients

eligible for multidisciplinary PRP. The experts were sent a link to an online survey. This survey

consisted of one open question in which the experts were asked to list all factors that might

infl uence case complexity.



6

Analysis

Duplicates of answers were removed and the remaining answers were clustered into comprehensive

factors as described in the International Classifi cation of Functioning, Disability and Health (ICF) 8. The fi nal clustering and disposition of the factors were consented on by three authors (FW, EB

and MR).

Second round

Objective: to identify the most important factors for determination of the case complexity. The

experts were sent a link to a second online survey, and they received the overview of the factors

arranged in the domains of the ICF model. The experts were asked to select the 20 most important

factors for the determination of the case complexity of patients eligible for a multidisciplinary PRP.

They were asked to select the factors in such a way that the factors comprised the complexity of

80% of the patients.

Analysis

Per factor the number of selections was counted. If possible, factors were clustered under higher

level factors in order to develop a manageable number of (comprehensive) factors for the third

round. The number of selections of all factors was calculated and the top 10 factors with the

highest sums were selected for the third round. Percentages of the total amount of selections per

factor were calculated. The factors were all linked to ICF categories using the ICF linking rules 9.

Third round

Objectives: to weigh the top 10 most selected factors and investigate how these factors are

assessed for a PRP. The experts were sent a link to a third online survey, which included the 10

factors with the highest sums of selections. This survey consisted of three questions:

Question 1: Please rate to what extent the factor accounts for the case complexity on a 5-point

Likert scale (0= no weight, 1= light, 2= average, 3= heavy, 4= very heavy).

Question 2: Please indicate how you determine the relevance of each individual factor (1=

based on measurement tool only (100%/0%); 2=more based on measurement tool than clinical

reasoning and examination (75%/25%); 3=based on measurement tool as much as on clinical

reasoning and examination (50%/50%); 4=more based on clinical reasoning and examination

than measurement tool (25%/75%); 5=only based on clinical reasoning and examination

(0%/100%)). An expert could also select the option ‘I do not determine this factor’.

Question 3a: If you based your assessment (partly) on measurement tools, please list the

measurement tools and the cut-off points you use. Question 3b: If you based your assessment

(partly) on clinical reasoning and examination, please explain how you determine the relevance of

the factor and which ‘cut-off points’ you use.

R1R2R3R4R5R6R7R8R9


96 | Chapter 6

Analysis

Question 1: For each factor, the mean, standard deviation, median and interquartile range (IQR) of

the ratings were calculated. The differences between the four different disciplines were analyzed

by means of the Kruskal-Wallis test. The differences between rehabilitation specialists and the

other disciplines as one group were analyzed by means of the Mann-Whitney U test. Question

2: medians and IQRs were calculated for each factor. Question 3: measurement tools and cut-off

points were listed.

Case Complexity Index

The mean ratings of relevance (Q1, 3rd round), were used to calculate the weights. The weights

were linearly transformed to obtain a CCI on a 0-100 scale. The formula for the complexity index

was formed by calculating the sum of the products of the factors and their mean weights. The

options for the scoring of every individual factor were 0 (not present), 0.5 (partially present) or

1 (fully present), or ‘cannot determine’. The minimum score of the presence of the factors was

0, calculated by scoring every factor 0. The maximum score was 10, calculated by scoring every

factor 1. The minimum total CCI score is 0 (no presence of any factor) and the maximum score is

100 (all factors are fully present).

Feasibility

A feasibility study of the CCI was performed at the University Medical Centre Groningen, Centre

for Rehabilitation. A convenience sample of 16 consecutive patients with CMP who underwent a

clinical assessment for a PRP, was included. The rehabilitation specialists, occupational therapists,

physiotherapists and psychologists acted as assessors by fi lling in a form following the clinical

assessments. The rehabilitation specialists fi lled out the form independently from the other

member of the team. The other members of the team fi lled out the form together during team

meetings after completion of the individual assessments. The form was designed to identify the

WPN level and the presence of the 10 factors of the CCI. The defi nitions of the WPN levels were

provided for this pilot study, each WPN level was divided into three categories: ‘-‘ (light), ‘normal’

(typical) or ‘+’ (severe). For the CCI, the assessors scored every factor to what extent the factor

was present, based the clinical judgement. The options were: ‘not present’ (score 0), ‘partially

present’ (0.5), ‘fully present’ (1.0) and ‘cannot determine’. The option ‘cannot determine’ was

considered as a missing value.

Analysis

CCI was calculated for every patient. To deal with the missing values, two methods were used:

(1) Fully Conditional Specifi cation Method in SPSS to replace the missing values (imputation), and

(2) dividing the complexity index by the number of fi lled in factors (mean complexity index). The

medians, means and standard deviations of the values of the factors, complexity index and WPN



6

level were calculated. Spearman’s rho was calculated between CCI and WPN level. To investigate

the relevance of the factors without the weights, Spearman’s rho was calculated between the

unweighted CCI (sum of the values of the factors for every patient) and the WPN level. The

analyses were conducted separately for the rehabilitation specialists and treatment team. The

interrater agreement between the rehabilitation specialists and treatment team on the WPN level

and the assessment of the factors was assessed by calculating Kappa with linear weighting.

Results

Participants

A total of 52 experts participated in the Delphi study: 21 rehabilitation specialists, 12 physical

therapists, 11 occupational therapists and 8 psychologists. Round 1: 44 experts; 17 rehabilitation

specialists, 11 occupational therapists, 10 physical therapists and 6 psychologists. Round 2: 41

experts (response rate= 93.2%); 16 rehabilitation specialists, 9 occupational therapists, 9 physical

therapists and 7 psychologists. Round 3: 36 experts (response rate= 81.8%); 13 rehabilitation

specialists, 9 occupational therapists, 8 physical therapists and 6 psychologists. The experts

worked in 15 rehabilitation centres in 11 out of 12 provinces in the Netherlands. Three non-

responding experts declared that lack of time was their reason for not responding.

First round

After removal of duplicates, the experts listed a total of 166 factors that were deemed important

for the determination of the case complexity (Appendix 1).

Second round

All experts selected 20 factors which they deemed most relevant for determining the case

complexity. Three experts selected less than 20 factors, which resulted in a total of 802 selections.

The majority of unique selections were clustered into 25 (comprehensive) factors in the second

round. The top 10 most selected factors are presented in table 1. These 10 factors represent

59.1% of the total amount of selections. Appendix 2 shows the defi nitions of the 10 factors.

R1R2R3R4R5R6R7R8R9


98 | Chapter 6

Table 1 Top 10 most important factors selected in the second round

Factor ICF domain ICF Category N selections % of total amount of selections

Psychiatric disorders ICD 123 15.3Motivation Body Functions b130 70 8.7Family system problems External Factors e310 47 5.9Mentalization capacity Body Functions b164 43 5.4Somatisation Body Functions b1602 42 5.2Financial and work-related problems External Factors e590 34 4.2Features of complaints Body Functions b280 34 4.2Treatment-interfering personality traits Body Functions b126 30 3.7Life events Personal Factors nc 27 3.4Personal injury proceeding External Factors e550 24 3.0Total selections / % 471 59.1

ICF: International Classifi cation of Functioning, Disability and Health; ICD: International Classifi cation of Diseases and Related Health Problems, Nc: not classifi ed

Third round

The results of the ratings of relevance of the individual factors are presented in table 2. The

mean ratings ranged from 1.8 for ‘features of complaints’ to 3.6 for ’psychiatric disorders’. The

mean ratings per discipline are not presented, because differences in medians between the four

disciplines were non-signifi cant. The rating of relevance per factor is shown in fi gure 1.

Table 2 Ratings of relevance of factors, differences between disciplines and between rehabilitation specialists and treatment team.

All disciplines RSa vs teamFactor Meanb SD Medianb IQR h (df=3)c p Ud pPsychiatric disorders 3.6 0.65 4 1 5.14 .162 144.0 .831Motivation 3.2 1.03 3.5 1 7.70 .053 74.5 .007Treatment-interfering personality traits 3.1 0.67 3 0 6.70 .082 145.5 .877Somatisation 2.8 0.75 3 1 3.14 .371 103.5 .099Family system problems 2.6 0.87 3 1 2.34 .505 119.5 .288Life events 2.4 0.84 2 1 4.12 .249 148.5 .972Mentalization capacity 2.4 0.87 2 1 4.15 .245 104.0 .107Personal injury proceeding 2.4 0.72 2 1 0.90 .826 129.0 .461Financial and work-related problems 2.3 0.69 2 1 0.16 .984 147.5 .941Features of complaints 1.8 0.84 2 1 3.17 .367 101.5 .090

a RS = Rehabilitation Specialistsb Options of 5-point Likert scale for importance: 0= no weight, 1= light, 2= average, 3= heavy, 4= very heavy.c h= test-statistic Kruskal Wallis test for differences between all disciplines. d U=test-statistic Mann-Whitney U test for differences between rehabilitation specialists and treatment team.



6

0 1 2 3 4

Psychiatric disorders

Motivation

Treatment-interfering personality traits

Somatization

Family system problems

Life events

Mentalization capacity

Personal injury proceeding

Financial and work-related problems

Features of complaints

Rehabilitation Specialists

Occupation therapists

Physical therapists

Psychologists

Total

Figure 1 Mean rating of importance of factors per discipline (0=no weight, 1=light, 2=average, 3=heavy, 4=very heavy)Factors are ordered by decrease of mean rating. Disciplines are divided and ordered by rehabilitation specialist and the rest of the team.

The results of the methods of assessment of the factors are shown in table 3. For all factors, the

medians were 3 or higher, indicating a tendency towards a dominance of clinical examination

and reasoning. In total 40 different instruments were named. The number of instruments that

are used to measure a factor ranged from 0 (life events) to 17 (treatment interfering personality

traits). In 49 cases cut-off points were not mentioned for an instrument used for a specifi c factor.

The most mentioned cut-off point for clinical examination and reasoning is ‘negative infl uence

of factor on functioning’.

R1R2R3R4R5R6R7R8R9


100 | Chapter 6

Table 3 The determination of the factors (n=36).

Factor Mediana IQR Missing (%)b

Psychiatric disorders 3 1 36.1Motivation 5 1 0.0Treatment-interfering personality traits 3 1 27.8Somatisation 4 1 5.6Family system problems 5 1 22.2Life events 5 1 22.2Mentalization capacity 5 0 25.0Personal injury proceeding 5 0 33.3Financial and work-related problems 5 1 30.6Features of complaints 4 2 8.3

a Options of 5-point Likert scale for determination: 1=only based on measurement tool, 2=more based on measurement tool than clinical reasoning, 3=based on measurement tool as much as on clinical reasoning, 4=more based on clinical reasoning than measurement tool and 5=only based on clinical reasoning. b The missing values represent the option ‘I do not determine this factor’

Case Complexity Index

The CCI is based on the 10 factors (round 2) and their weights (0-4; 3rd round). The weights

were linearly transformed to a 0-100 scale by multiplying the mean ratings by 3.79 (= 100/

sum weights). The fi nal formula for the CCI = 13.5*psychiatric disorders + 12*motivation +

11.6* treatment-interfering personality traits + 10.6*somatisation + 10*family system problems

+ 9.2*life events + 9.1*mentalization capacity + 8.9*personal injury proceeding + 8.5*fi nancial

and work-related problems + 6.6*features of complaints.

The options for the scoring of every factor are 0 (not present), 0.5 (partially present) or 1 (fully

present), or ‘cannot determine’ (missing value). The minimum score is 0 (no presence of any

factor) and the maximum score is 100 (all factors are fully present). A higher CCI indicates a more

complex case. The factors can be scored on a CCI form (table 4). For example, a fi ctive patient

scored 1 on motivation, mentalization capacity and fi nancial and work-related problems, and

scored 0.5 on family system problems, somatisation, treatment-interfering personality traits and

life events. The scores were multiplied with their corresponding weights, summed and CCI of this

patient is 50.3.



6

Table 4 Example of the CCI for a fi ctive patient.

Optionsa

Score WeightScore* weight

Not present

Partially present

Fully present

Cannot determine

Psychiatric disorders 0 13.5 0.0Motivation 1 12.0 12.0Treatment-interfering personality traits 0.5 11.6 5.8Somatisation 0.5 10.6 5.3Family system problems 0.5 10.0 5.0Life events 0.5 9.2 4.6Mentalization capacity 1 9.1 9.1Personal injury proceeding 0 8.9 0.0Financial and work-related problems 1 8.5 8.5Features of complaints 0 6.6 0.0CCI 50.3

a Not present=0, Partially present=0.5, Fully present=1, Cannot determine=missing value. CCI=Case Complexity Index

Feasibility

In the feasibility study 16 patients were assessed. Fourteen patients were assessed by the

treatment team and, 14 patients were assessed by the rehabilitation specialists. Twelve patients

were assessed by both the rehabilitation specialists and treatment team. The patients had a mean

age of 41.2 years (SD=15.1), 7 were female, and 8 patients were diagnosed with lumbago, 6

with wide spread pain and 2 with chronic pain syndrome. The mean Pain Disability Index (PDI) 10 score of the patients was 39.5 (SD=11.5, range: 16 to 56). The medians, means and standard

deviations of the values of the factors, CCI and WPN level are shown in table 5. The correlations

between CCI and WPN were strong and signifi cant for the treatment team, lower and non-

signifi cant for the rehabilitation specialists. Table 6 shows both methods used for missing values:

unweighted CCI, (mean CCI was used for missing values), and CCI (missing values were imputed).

The agreement of the assessment of the WPN level between the rehabilitation specialists and the

treatment team was non-signifi cant (k=0.17, 95% CI 0.000 to 0.402). The agreement of the

assessment of the CCI between the rehabilitation specialists and the treatment team was k=0.48

(95% CI 0.355 to 0.607), which shows a moderate agreement 11. All assessors were able to

comprehend the CCI form and its concepts with brief written instructions only. The time needed

to fi ll out the CCI was less than 1 minute per patient.

R1R2R3R4R5R6R7R8R9


102 | Chapter 6

Table 5 Descriptives of the scores on the factors of the CCI and WPN level in the feasibility study

Rehabilitation specialists (n=14a)

Treatment team (n=14a)

Median Mean SD Median Mean SDPsychiatric disorders 0.25 0.25 0.26 0.00 0.31 0.43Motivation 0.00 0.21 0.38 0.00 0.25 0.38Treatment-interfering personality traits 0.50 0.50 0.28 0.50 0.61 0.35Somatisation 0.50 0.36 0.36 0.50 0.43 0.43Family system problems 0.50 0.46 0.24 0.50 0.46 0.37Life events 0.50 0.54 0.37 0.50 0.65 0.38Mentalization capacity 0.00 0.18 0.32 0.00 0.29 0.38Personal injury proceeding 0.00 0.04 0.13 0.00 0.04 0.13Financial and work-related problems 0.50 0.54 0.43 0.50 0.50 0.44Features of complaints 1.00 0.86 0.23 1.00 0.71 0.38 CCI (missing values imputed) 37.32 16.33 41.82 19.76WPN level 3/3+ 3+

CCI=Case Complexity Index

Table 6 Correlations between complexity and WPN level using different methods CCI.

WPN levelRehabilitation specialists Treatment team

r p r pUnweighted CCI 0.45 0.146 0.73 0.003*CCI (missing values imputed) 0.43 0.134 0.72 0.003*

WPN level: r= spearman’s rho,CCI=Case Complexity Index, * p<0.05

Discussion

This study explored the most important factors to assess case complexity of patients with CMP

eligible for PRP, according to the experts in the Dutch chronic pain rehabilitation fi eld. These factors

were used to initiate the development of a CCI. The CCI consists of 10 factors: fi nancial and work-

related problems, features of complaints, personal injury proceeding, life events, mentalization

capacity, motivation, personal injury proceeding, psychiatric disorders, somatisation and family

system problems. Of these 10 factors, according to the experts, psychiatric disorders have the

most infl uence on the case complexity and features of complaints have the least infl uence. The

four professions agreed on the ratings of infl uence of the factors. The assessment of the factors

is mainly based on clinical examination and reasoning. Feasibility of the CCI was established.

The factors in our study are mainly psychosocial factors. Indicating that the experts rate the

biological factors to be of lower relevance in the assessment of the case complexity of patients

eligible for PRP. This is remarkable considering PRPs are based on the biopsychosocial model. The



6

results of this study showed that there are no unambiguous measurement tools for the factors,

which is in line with a systematic review that there is a large variability in measurement tools for

predictors of outcome of multidisciplinary PRPs for patients with chronic low back pain 12. These

predictors overlapped partially with the factors in the current study.

The CCI is based on a combination of 10 factors, which is a number chosen by the authors to

construct a manageable tool. These factors account for 59.1% of the total number of selections.

Therefore, it does not capture all factors that determine case complexity, but it does contain the

most important factors as viewed by the experts. Also, the case complexity does not take the

interaction between the factors into account. The goal of this Delphi study was to gain insight

into the assessment of the case complexity based on opinions of experts and take fi rst steps to

create a new concept for the assessment of case complexity based on this insight. Therefore, the

results of this Delphi study do not represent “the truth”. It presents an expert view about factors

of case complexity, and a state of the art on how these factors are assessed in rehabilitation

practice. The experts were all from the Netherlands, which may limit generalizations to pain

rehabilitation settings in other countries.

The CCI might be a useful tool for the assessment of the case complexity of patients with CMP.

It is transparent, easy to score, takes very little time to fi ll out and provides a numeric value

to the case complexity. The complexity index can be used by the rehabilitation specialists and

other members of the treatment team. For the assessment of the treatment team, the results

of the pilot study show that the CCI is associated with the WPN level. The CCI is currently still a

rather subjective measure, just as the gold standard WPN method, because the decision on the

presence of each factor is based on clinical reasoning. However the criteria of the CCI are more

specifi c, which can provide a basis for structured assessment of case complexity and stimulation

to validate the decision on assessing each factor. Other advantages of the CCI compared to the

WPN level is the better differentiation between patients, because of its numeric value (0-100)

compared to (broad) categories and the possibility to convert the factors into treatment goals.

On the limitations side, contra-indications for the treatment were not considered. Factors may

be too dominantly present that it cannot be treated with the expertise of a typical PRP team. For

example, a patient with a dominantly present personality disorder that interferes with PRP may

not benefi t from a PRP, and may be better referred to other specialists. Also one factor can be

heavily present, needed much therapy, but need other therapy than multidisciplinary PRP. The

present CCI does not account for this.

Some other attempts to capture complexity have been described the Rehabilitation Complexity

Scale 13, the Vector Model of Complexity 14, COMPRI 15 and INTERMED 16. However none of these

methods are designed to assess case complexity specifi cally for patients with CMP. As stated 17,

R1R2R3R4R5R6R7R8R9


104 | Chapter 6

there is need for more research to assess case complexity for more specialist rehabilitation

programs for individual patients within the rehabilitation process. The CCI is a more specifi c

measure, only addressing relevant factors for the clinical assessment of patients with CMP who

are eligible for a multidisciplinary PRP and includes factors derived from experts in the fi eld.

Therefore, the factors of the CCI and its score seem to be more appropriate for the use within

this specifi c setting of multidisciplinary PRP.

Our study showed that the assessment is mainly based on clinical examination and reasoning.

Also, the used cut-off points for the clinical examination and assessment are highly variable for

each factor. Because the clinical examination and reasoning will be infl uenced by the assessors’

experiences and beliefs and various implicit cut-off points are used, it is challenging to reach

consensus about the case complexity of an individual patient. In this way, the clinical examination

and reasoning of the assessor will likely infl uence the determined case complexity of a patient.

This can still be the problem with the CCI, which showed a fair interrater agreement between

the rehabilitation specialists and treatment team. However, these results are based on a small

sample size and many different assessors were involved. The defi nitions of the factors were also

not given to the assessors, which may have lead to variability in the interpretation of the factors.

A few assessors were also involved in the Delphi study and therefore had more knowledge of

the factors. These exploratory analyses were performed within the context of feasibility. Clearly,

formal studies are needed to study psychometric properties of the CCI.

The fi rst steps of the development of the CCI have been reported in this paper. However, we

realize that the present CCI is not an end-product. The contribution of each factor to the case

complexity should be tested in a sensitivity analysis. In this analysis other factors should also be

considered. To improve the CCI, it is recommended to objectify the assessment of the individual

factors. For some factors validated instruments are available and these could be incorporated

in the CCI. It might be necessary to have an extra option between ‘partially present’ and ‘fully

present’ for the scoring of the presence of the factors (e.g. similar to the ICF qualifi ers), as there

is a large gap between those options. Because the present CCI represents 59.1% of the case

complexity, the addition of an extra open factor may be considered to facilitate for the 40.9%

that are unaccounted for. Contra-indications for the treatment should be explored and added to

the CCI. The validity, reliability and the practicality of the CCI need to be tested in several settings

to ensure the generalizability within chronic pain rehabilitation. Future research is also needed to

explore the usability of the total CCI score compared to the scores and weights per factor and

analyse relationships of case complexity with dosage of PRPs.



6

Acknowledgements

The authors acknowledge the participation of all experts in the Delphi study and the participation

of the rehabilitation physicians and the other team members in the feasibility study.

R1R2R3R4R5R6R7R8R9


106 | Chapter 6

References

1. Gatchel RJ, Peng YB, Peters ML, Fuchs PN, Turk DC. The biopsychosocial approach to chronic pain: Scientifi c advances and future directions. Psychol Bull. 2007;133(4):581-624.

2. Dansie EJ, Turk DC. Assessment of patients with chronic pain. Br J Anaesth. 2013;111(1):19-25.




6. Boonstra AM, Schiphorst Preuper HR, Brouwer MAH, Weerdt de M, Knol D, Köke A. Classifi catie van patiënten met chronische pijnklachten: Onderzoek naar de betrouwbaarheid van de WPN niveaus. Ned T Pijnbestrijding. 2011;30:5-10.

7. Delbecq AL, Van de Ven AH, Gustafson DH. Group techniques for program planning: A guide to nominal group and delphi processes. Glenview, Ill: Scott, Foresman; 1975:174.

8. World Health Organization. ICF: International classifi cation of functioning, disability and health. Geneva: World Health Organization; 2001:299.

9. Cieza A, Geyh S, Chatterji S, Kostanjsek N, Ustun B, Stucki G. ICF linking rules: An update based on lessons learned. J Rehabil Med. 2005;37(4):212-218.


11. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159-174.

12. van der Hulst M, Vollenbroek-Hutten MM, Ijzerman MJ. A systematic review of sociodemographic, physical, and psychological predictors of multidisciplinary rehabilitation-or, back school treatment outcome in patients with chronic low back pain. Spine (Phila Pa 1976). 2005;30(7):813-825.

13. Turner-Stokes L, Williams H, Siegert RJ. The rehabilitation complexity scale version 2: A clinimetric evaluation in patients with severe complex neurodisability. J Neurol Neurosurg Psychiatry. 2010;81(2):146-153.

14. Safford MM, Allison JJ, Kiefe CI. Patient complexity: More than comorbidity. the vector model of complexity. J Gen Intern Med. 2007;22 Suppl 3:382-390.

15. Huyse FJ, de Jonge P, Slaets JP, et al. COMPRI--an instrument to detect patients with complex care needs: Results from a European study. Psychosomatics. 2001;42(3):222-228.

16. de Jonge P, Huyse FJ, Slaets JP, Sollner W, Stiefel FC. Operationalization of biopsychosocial case complexity in general health care: The INTERMED project. Aust N Z J Psychiatry. 2005;39(9):795-799.

17. Wade D. Complexity, case-mix and rehabilitation: The importance of a holistic model of illness. Clin Rehabil. 2011;25(5):387-395.



6

R1R2R3R4R5R6R7R8R9


110 | Chapter 7


General discussion | 111

7

The main aim of this thesis was to acquire insight into aspects of dosage for multidisciplinary

Pain Rehabilitation Programs (PRPs) for patients with chronic pain. Five studies were performed.

In this concluding chapter, the main fi ndings and factors related to PRP dosage will be discussed,

followed by clinical, scientifi c and societal implications, strengths and limitations of this thesis and

recommendations for future research.

Main Findings

A systematic review and meta-analysis (Chapter 2) revealed that PRP dosage had never been studied

as a primary objective within pain rehabilitation research. Analyzing studies that investigate the

effects of PRPs and comparing the reported dose variables resulted in an overview of PRPs with

a wide variety of dosages having different outcome measures and effects on disability, quality of

life (QoL) and work participation. Evaluation moment, number of disciplines, type of intervention,

duration of intervention, percentage of women, and age infl uenced the effect of PRP. However,

because dose variables and content variables were strongly interrelated, a distinction between

these variables regarding their effects was impossible.

Chapter 3 describes a clinical study in which the course of disability reduction during PRP was

analyzed. At a group level, the course of disability reduction occurred following a quadratic time

model. This course was infl uenced by initial Pain Disability Index (PDI) scores, treatment week,

average pain, and the interaction between initial PDI scores and treatment week, resulting in, for

example, a steeper slope (faster disability reduction) for patients with higher initial PDI scores and

average pain.

Chapter 4 describes a RCT in which the effects of PRPs with similar content but different dosages

were compared. Frequent extension of the Care as Usual – Short Form (CAU-SF) program led

to exceeding predefi ned stopping rules, and the study had to be stopped prematurely. Because

of lack of power and extension of CAU-SF, differences in dosages between the control and

experimental group were too small to detect signifi cance of non-inferiority. Hence, no robust

conclusions could be drawn. Ancillary regression analysis showed that dose variables did not

contribute to the prediction model, which could be interpreted as a possibility of shortening PRP

without loss of effect.

Chapter 5 describes a study in which patient and therapists’ perceptions were explored regarding

dosages of PRP in three rehabilitation centers in the Netherlands with different dosage of PRP.

Factors that determined the dosage were related to patient, treatment and external factors. The

results showed that although PRPs differed in dosage, perceptions of factors that were important

R1R2R3R4R5R6R7R8R9


112 | Chapter 7

for determining dosage were similar. Remarkably, similar factors seemed to lead to different

choices in terms of PRP dosage in different centers.

Chapter 6 describes a study in which factors related to the general status, defi ned as case

complexity of patients eligible for PRP, were explored. The 10 most important factors were listed

and weights of infl uence were explored, resulting in a case complexity index (CCI), which was

tested for feasibility in clinical practice. Our hypothesis of a possible relationship between case

complexity and dosage was consistent with the theme of “general status” from the interviews

described in Chapter 5. Patients and rehabilitation professionals mentioned a relationship between

dosage and the general status of a patient, which includes factors similar to the factors in the

CCI. Therefore, it could be assumed that the CCI could be tested in clinical practice to determine

case complexity and to relate this to the prediction of PRP dosage. The relationship between case

complexity and dosage should be tested to confi rm or reject this hypothesis.

Factors associated with PRP dosage

While these fi ve studies have not led to defi nitive conclusions about the optimum PRP dosage, it

has contributed relevant insights regarding factors related to dosage and the effects of PRP. These

can be divided into 1: factors related to the effect of PRP; 2: patient related factors; 3: PRP related

factors; 4: common therapeutic factors and, 5: external factors.

Factors related to the effect of PRP

Literature and studies performed in this thesis showed a wide variety in PRP dosage 1,2. To determine

the optimum PRP dosage, the question of “how much therapy is needed to establish the desired

effect” should be answered. However, up until now, most studies have analyzed effects of PRP

at a certain dosage by analyzing pre/post and follow-up outcomes 3-6, not knowing whether this

effect could also be reached with a lower dosage or a greater effect could be reached at a higher

dosage. Generally, studies do not explicitly target a certain effect, nor do they explicitly adjust

dosages towards this target effect. Additionally, the question arises regarding which effect should

be analyzed with respect to PRP: the effect of PRP on improving participation or the effect patients

reach at a certain moment using better self-management skills acquired from PRP. For example,

can PRP considered effective if a patient does not acquire a clinically important change on PDI,

but shows good skills to self-manage his disabilities in the future? Moreover, multidimensional

effects of multidisciplinary therapies, including psychological components, are more diffi cult to

defi ne, compared to the unidimensional effect of mono-disciplinary therapy, including mainly

physical components. Additionally, it is easier to measure “physical condition” compared to

measuring ”the ability to improve physical condition”. The current literature shows a wide variety



7

of outcome measures and domains addressing the effect of PRP in relation to participation, such

as disability reduction, QoL and work participation 1,2,7. However, other outcome measures related

to self-management skills should be added to assess the effect of PRP and could be used to assist

with determining the optimum dosage. Being able to self-manage pain-related disabilities would

presumably effect participation in daily life 8.Measuring self-management skills as an effect of PRP,

in addition to rehabilitation programs in general, is in line with changes in the new concepts of

health. Instead of describing health as “a state of complete physical, mental and social wellbeing

and not merely the absence of disease or infi rmity” 9, Huber et al. stated that health is “the ability

to adapt and to self-manage, in the face of social, physical and emotional challenges” 10.

Regarding the effects of PRP, it can be debated from which perspective the optimum effect

should be targeted – optimum from the perspective of the patient, from the rehabilitation

professionals or from an economic perspective. In the current system of healthcare fi nance in

the Netherlands, and in some cases, from an economic perspective, a higher PRP dosage may

provide more income for the rehabilitation provider, while from the perspective of a patient and/

or rehabilitation professional, the desired effect may have already been reached. If the optimum

dosage is defi ned from this economic perspective, overtreatment could occur. For these reasons,

more insight is needed regarding which factors are important to determine optimum PRP dosage.

Additionally, scientifi c evidence and consensus regarding target effects of PRP are needed to

improve comparability and the ability to determine optimum PRP dosage.

Patient related factors

Dosage of PRP could be related to the patients’ case complexity at the beginning and during PRP.

Chapters 2, 3, 5 and 6 showed that several patient characteristics and patient-related factors,

both physical and psychological, can, or are assumed to, infl uence the effect of PRP and the

required dosage. Taking together the fi ndings of this thesis and the questions regarding the

effects of PRP, there could be an association between the stages of change 11 and PRP dosage.

The transtheoretical model of changing behavior presumes that the readiness of change can be

described in different stages of change. During the precontemplation stage, patients believe that

chronic pain is primarily a biomedical problem and rehabilitation professionals could help relieve

the pain during PRP 12. Therefore, they are not willing to change behavior and are not motivated to

take the biopsychosocial approach offered in PRP. At the contemplation stage, patients are aware

of the biopsychosocial nature of their problem, but do not know how to manage this. Therefore

at this stage, a patient needs professionals help to manage the pain and related disabilities. In the

later stages (preparation and action stages), patients have learned how to self-manage their pain

and disabilities and they need to maintain this healthy behavior. Being able to self-manage pain-

related disabilities and to make plans for maintaining healthy behaviors will theoretically result in

less need for professional help, and therefore, less, or even no, PRP. Self-management skills will

R1R2R3R4R5R6R7R8R9


114 | Chapter 7

theoretically also result in late-treatment improvements in outcomes of reduction of experienced

disability, QoL and work participation. This is in line with the results from previous studies 12,13.

A recent study 14 analyzed changes in precontemplation and action attitudes during cognitive-

behavioral therapy (CBT) for patients with chronic pain. From this study, it was concluded that

there might be an association between changes in the stages of behavior modifi cation and the

following outcome improvements. However, at present, this association has not been proven

regarding PRPs, there has been no documented association related to PRP dosage. In addition,

measuring the stage of change of a patient at a certain moment is still diffi cult and arbitrary 14.

PRP related factors

PRP itself, as well as the rehabilitation professionals involved in PRP, could infl uence dosage. PRP

is offered at a certain dosage – total duration in weeks, total number of contact hours, number

of rehabilitation professionals involved, form of intervention and physical dosage of exercise

(weight, repetition, duration, frequencies). These dosage variables differ across PRPs, nationally

and internationally. From the systematic review and the interviews conducted in Chapters 2

and 5 respectively, it can be concluded that there is a lack of scientifi c evidence regarding on

which factors the offered PRP dosage is based. Choices of dosages of PRPs included in Chapter

5 are mainly based on clinical expertise and historical grounds. Remarkably, both rehabilitation

professionals and patients are generally satisfi ed with the offered and received PRP, regardless of

the large differences in offered dosages. Patient satisfaction was studied in a systematic review

within physical therapy care 15, concluding that therapist characteristics, the process of care, and

the organization of care were consistently identifi ed factors that infl uenced satisfaction with care,

whereby treatment outcome and expectations of physical therapy were less-frequently reported

factors.

Dosage can be expressed in terms of total duration, in addition to expressing it in terms of total

number of contact hours. For example, PRP can be offered for the same number of weeks (total

duration), but with a greater or fewer total number of contact hours. These differences in the total

duration and corresponding number of contact hours per week are based on different choices

regarding the content of contact hours, the number of rehabilitation professionals, and the form

of PRP. The choice of content of contact hours, for example, is based on the assumption of what

content should be offered within the contact hours of PRP and which activities or homework should

be performed by the patient himself outside the rehabilitation center. Apparently, rehabilitation

teams have made different choices regarding offered content and related dosages based on

different assumptions and the best available evidence 1. Several studies have been performed

to analyze components of PRP, such as differences between outpatient and inpatient PRP 16 and

graded activity versus graded exposure 17. However, evidence for the optimum composition of

PRP and related optimum PRP dosages, including all components and factors, is lacking.



7

Common factors

Common therapeutic factors 18 are described as possible explanations for equivalent effectiveness

across different treatment approaches within psychotherapy. Rosenzweig wondered whether

“the factors alleged to be operating in a given therapy are identical with the factors that

actually are operating, and whether the factors that actually are operating in several different

therapies may not have much more in common than have the factors alleged to be operating”.

Examples of common factors are reassurance, feedback, insight, success experiences, therapeutic

alliances, patient expectations, therapist expertness, and trust 19. Because of the biopsychosocial

characteristics of PRP, common factors could probably infl uence the effects and PRP dosage. The

results of the qualitative study in Chapter 5 showed that these common factors are similar but

lead to different choices in PRP dosage. However, little evidence is available regarding specifi c

and common factors and their relationship to outcomes of chronic pain treatment 14,20-22 and its

dosage.

Overall, this thesis opened a black box in pain rehabilitation research. It is the fi rst step in

unraveling the complexities of PRP dosages. Still, it is a huge challenge to answer the question of

optimum PRP dosage, taking into account the relationship of dosage and content, including all

of the above-mentioned factors. Adding the factor of dosage to what Burns 14 stated regarding

the working mechanisms for CBT, “it is not yet clear whether CBT worked for reasons specifi ed by

theory, through the inspiration of hope and the support and encouragement of a good therapist,

or a combination of factors”.

External factors

The legislation of expenses and reimbursements in healthcare could lead to differences in PRP

dosages across countries. Direct and indirect costs become more and more important. In general,

PRPs have proven cost effectiveness 23,24, but it is a challenge to balance how much treatment is

needed versus its costs. Because the effect of PRP is not strictly defi ned, costs of PRP can be the

motivation for offering a certain PRP dosage in order to ensure cost benefi ts. In 2014, the Dutch

government started, in cooperation with the Dutch Association of Physicians in Rehabilitation

Medicine, to develop a new system of fi nancing for rehabilitation medicine. Physicians across

the Netherlands described the content of rehabilitation treatment according to different therapy

components. Registering contact hours linked to these therapy components will lead to more

transparency for all rehabilitation activities performed during rehabilitation programs across the

Netherlands. Based on the combination of content and contact hours, a new fi nance system will

be implemented before 2019. The present fi nance system is only based on the number of contact

hours divided across categories, regardless of the content or effect of therapy. The adjustment

of content in terms of therapy components related to contact hours of rehabilitation programs

will probably draw attention and discussion regarding dosage of programs because of the link to

R1R2R3R4R5R6R7R8R9


116 | Chapter 7

funding. Consequently, if dosages will be adapted, this will lead to more comparability of these

programs across the Netherlands.

This thesis showed that PRP dosage is associated with many implicit and un- or under-researched

factors. Therefore, it is presently diffi cult to draw robust conclusions about the optimum PRP

dosage, and further research is needed to acquire more insight into, and scientifi c evidence of,

differences and similarities between these factors and their interrelationships.

Implications

Clinical implications

In general, multidisciplinary PRPs have proven their effectiveness 1,6,7,25. However, no distinction

can be made between different PRPs. This thesis resulted in a critical look at the different aspects

of dosage, and it should encourage clinical practice and research to focus more on PRP dosage.

Implications for patients

Patients should be made aware of the different dosages. In the long term, research aimed at

optimizing PRP dosage will benefi t patients because healthcare providers will be better able to

estimate which PRP dosage is required for subgroups of patients. Eventually, patients will receive

the most effective and effi cient PRP.

Implications for therapists

Therapists involved in PRP should refl ect on the dosage of the offered program. They have to

be aware of the differences of dosage within PRPs and they should be aware of the distinction

between what is “nice to have or need to have”. Even though it seems like there is always

something to practice or teach, there is room for improvement and patients have the right to

receive the optimum care related to content and dosage of treatment. Over- or under-dosage of

treatment will not benefi t anyone. Therapists should integrate all factors related to the patients’

needs and preferences, as well as their professional experience, and they should be aware of the

infl uence of common factors, costs and the available evidence to offer optimum PRP dosage.

Implications for rehabilitation centers

Insight into PRP dosage and being able to offer optimum PRP dosage will benefi t the entire

rehabilitation process. It will benefi t the patient fl ow, and will reduce waiting lists and over- or

underdosing treatment. As a result, centers will be able to treat the maximum number of patients

suffering from chronic pain and offer cost-effective PRPs. The issue of dosage studied in this thesis,

while studied within the pain rehabilitation context, is not unique for pain rehabilitation only. The



7

results of this thesis can be used as an example for other rehabilitation programs. Rehabilitation

programs assist in improving health in terms of “the ability to adapt and self-manage,” and all

rehabilitation programs can use this concept to offer more cost-effective programs.

Scientifi c implications

The results of this thesis, combined with the present developments regarding costs of healthcare,

indicate that dosage of rehabilitation programs is important to include in analyzing these

programs. Dosage of PRP has been a neglected topic in pain rehabilitation research, as well

as in rehabilitation research in general. Previous studies analyzing the effect of PRPs could

be confounded, and therefore, conclusions should be interpreted with caution. A theoretical

example: a study comparing the effectiveness of physical therapy to PRP. The total duration of

both programs is 10 weeks; however, physical therapy is given two hours a week and PRP is given

10 hours a week. PRP showed superior results, but it is questionable whether the content or the

dosage is superior, or whether other common factors were responsible for differences in the

effects. More research is needed to analyze what and how much works for whom.

Societal implications

Societal implications in general

The results of this thesis contribute to the ability to optimize PRP dosage and other rehabilitation

programs in general. Obtaining evidence of optimum dosage will lead to the reduction of over-

or underdosing treatment, thereby preventing the overdosage of treatment that leads to more

“needs”. Over- or underdosing also leads to the opposite of self-management.

Implications for (work)participation

The effect of patients being able to self-manage their pain will subsequently optimize participation

in daily life. Insight into “how much works for whom” will lead to the most desirable effects at

the optimum PRP dosage for each individual. This will allow them to participate in daily life and

return to work as soon as possible.

Financial implications

Because this thesis is an initial attempt at unraveling the complexities related to PRP dosage,

at this stage, no fi nancial implications can be described since there is no solid evidence on the

dose–effect relationship.

R1R2R3R4R5R6R7R8R9


118 | Chapter 7

Future research

Future studies in rehabilitation research should report dosage of treatment. This will improve

transparency, comparability and generalizability of rehabilitation treatment dosages and will give

researchers the ability to analyze dose–effect relationships. Within pain rehabilitation research,

future studies should also focus on the distinction between desired effect of PRP itself to be able

to fi nish PRP and the desired long-term effect of the patient self-management skills regarding

disability, QoL and (work) participation. Dosage of PRP should be analyzed together with all

specifi c and aspecifi c factors, such as a core set of working mechanisms of effective PRPs. In the

beginning, analyzing optimum PRP dosages appeared to be diffi cult. However, this thesis has

yielded an overview of dosage-related factors. The results can be combined into new hypotheses

regarding PRP dosages. The course of disability, which is reduced over time, could be related to

the desired effect and the stages of change and needs of patients during PRP. This qualitative

study confi rmed the relationship between case complexity and dosage. The Delphi study gives

substance to case complexity with the CCI, which can be used in clinical practice in relation to the

required dosage, and can also be used to scientifi cally analyze this relation. Moreover, the RCT

provides evidence that both arms with different dosages resulted in positive effects. This coincides

with the positive experiences of the PRP at different dosages used in the qualitative study, as

well as with the overview of effect sizes in the systematic review. These insights could be further

analyzed to determine whether there is a trend towards shortening PRPs. Future research could

result in more scientifi c evidence for optimum dosage, which will lead to better insight into what

works for whom and how much therapy is needed. This will accomplish the optimum effects at

the optimum dosage, with an addition focus on cost-effectiveness in terms of direct and indirect

healthcare costs.

Strengths and limitations

The research questions from this thesis evolved out of clinical practices, resulting in strengths

and limitations since the methodological strictness is not always compatible with clinical practice.

Because the studies in this thesis were the fi rst that analyzed PRP dosage, and no evidence was

available beforehand mistakes were made. Some choices made became weaknesses. Most studies,

except the systematic review, were performed in the Netherlands within the context of the Dutch

healthcare system, which limits generalization. However, although the Dutch healthcare system

differs from the healthcare systems in other countries, PRP dosage and rehabilitation in general is

an important issue all over the world. This thesis could be used to perform similar studies within

different countries to optimize dosage of therapy in each particular country. A strength of the

thesis is the variety of methods used. Qualitative and quantitative methods were used to reveal



7

different factors related to PRP dosage, resulting in an overview of important factors that now

need to be analyzed in future research to unravel and improve optimum PRP dosages.

Conclusion

This thesis is the fi rst that focuses on PRP dosage. Consequently, it should be regarded as a

starting point for determining factors related to dosages and exploring optimum PRP dosages.

This thesis analyzed dosage of PRP using different research methods and acquired different

insights in aspects of dosage. Therefore, this thesis contributes to the body of research in the area

of pain rehabilitation research aimed at analyzing the effects of all components and their dosages.

Additionally, results from the studies within this thesis contribute to societal developments in

transparency, comparability and cost effectiveness of PRPs and can be a basis for others to analyze

the dosages of rehabilitation programs.

R1R2R3R4R5R6R7R8R9


120 | Chapter 7

References



3. McCracken LM, Turk DC. Behavioral and cognitive-behavioral treatment for chronic pain: Outcome, predictors of outcome, and treatment process. Spine (Phila Pa 1976). 2002;27(22):2564-2573.


5. Gatchel RJ, Okifuji A. Evidence-based scientifi c data documenting the treatment and cost-effectiveness of comprehensive pain programs for chronic nonmalignant pain. J Pain. 2006;7(11):779-793.



8. Jensen MP, Nielson WR, Turner JA, Romano JM, Hill ML. Changes in readiness to self-manage pain are associated with improvement in multidisciplinary pain treatment and pain coping. Pain. 2004;111(1-2):84-95.

9. WHO. http://apps.who.int/gb/bd/PDF/bd47/EN/constitution-en.pdf?ua=1&ua=1. Updated 20052015.

10. Huber M. Towards a new, dynamic concept of health. its operationalisation and use in public health and healthcare, and in evaluating health effects of food.; 2014.

11. Prochaska JO, DiClemente CC. Transtheoretical therapy: Toward a more integrative model of change. Psychotherapy: Theory, Research & Practice. 1982;19(3):276-288.

12. Burns JW, Glenn B, Lofl and K, Bruehl S, Harden RN. Stages of change in readiness to adopt a self-management approach to chronic pain: The moderating role of early-treatment stage progression in predicting outcome. Pain. 2005;115(3):322-331.

13. Glenn B, Burns JW. Pain self-management in the process and outcome of multidisciplinary treatment of chronic pain: Evaluation of a stage of change model. J Behav Med. 2003;26(5):417-433.

14. Burns JW, Nielson WR, Jensen MP, Heapy A, Czlapinski R, Kerns RD. Specifi c and general therapeutic mechanisms in cognitive behavioral treatment of chronic pain. J Consult Clin Psychol. 2015;83(1):1-11.

15. Hush JM, Cameron K, Mackey M. Patient satisfaction with musculoskeletal physical therapy care: A systematic review. Phys Ther. 2011;91(1):25-36.

16. Boonstra AM, Reneman MF, Schiphorst Preuper HR, Waaksma BR, Stewart RE. Differences between patients with chronic musculoskeletal pain treated in an inpatient or an outpatient multidisciplinary rehabilitation program. Int J Rehabil Res. 2014;37(2):187-191.

17. Leeuw M, Goossens ME, van Breukelen GJ, et al. Exposure in vivo versus operant graded activity in chronic low back pain patients: Results of a randomized controlled trial. Pain. 2008;138(1):192-207.



7

18. Rosenzweig S. Some implicit common factors in diverse methods of psychotherapy. Journal of Psychotherapy Integration. 2002;12(1):5-9.

19. Lambert MJ, Ogles BM. Common factors: Post hoc explanation or empirically based therapy approach? Psychotherapy. 2014;51(4):500-504.

20. Thorn BE, Burns JW. Common and specifi c treatment mechanisms in psychosocial pain interventions: The need for a new research agenda. Pain. 2011;152(4):705-706.

21. Jensen MP. Psychosocial approaches to pain management: An organizational framework. Pain. 2011;152(4):717-725.

22. Day MA, Jensen MP, Ehde DM, Thorn BE. Toward a theoretical model for mindfulness-based pain management. The Journal of Pain. 2014;15(7):691-703.

23. Turk DC. Clinical effectiveness and cost-effectiveness of treatments for patients with chronic pain. Clin J Pain. 2002;18(6):355-365.

24. Sletten CD, Kurklinsky S, Chinburapa V, Ghazi S. Economic analysis of a comprehensive pain rehabilitation program: A collaboration between fl orida blue and mayo clinic fl orida. Pain Med. 2015;16(5):898-904.


R1R2R3R4R5R6R7R8R9


124 | Summary


Summary | 125

S

Pain rehabilitation programs (PRPs) consist of multidisciplinary treatments based on the

biopsychosocial model. It can assist people with chronic musculoskeletal pain to manage their

pain more effectively, and it can decrease pain related disabilities. Scientifi c research has shown

that pain rehabilitation has a positive effect on disabilities, work participation and quality of life.

However, there is a large amount of variation with regard to the content, form (inpatient or

outpatient, individual or in a group), duration and intensity of treatments, and the number of

disciplines involved in PRPs. In literature, very little is known about dosage of PRPs. Therefore,

the aim of the studies in this thesis was to gain insight into the effect of the different aspects of

dosage of PRPs. Various methods were used during the research and are described in the chapters

of this thesis. In this thesis, dosage of PRP has been defi ned as the total duration of treatment in

weeks, the total amount of contact hours, and the number of contact hours per week (intensity).

In Chapter 2 a systematic review is described. During the review process, no studies were found

that primarily dealt with the analysis of dosage of PRPs. Therefore, the aim of this review was

to study randomized controlled trials that analyzed the effects of PRPs. A requirement was that

these studies included dosage of treatments. The result is an overview of these studies, dosage of

the treatments that were compared and the accompanying effects sizes regarding the domains:

disabilities, work participation, and quality of life. Dosage of PRPs varied from 6.4 to 196.8

contact hours. In addition, the infl uence of dosage on the effect sizes was analyzed. A number of

variables were found to infl uence effects. Because these variables were strongly related to dosage

and it was not possible to isolate other factors, no robust conclusions could be drawn with regard

to the infl uence of dosage on the effects.

Chapter 3 presents a prospective cohort study. Its objective was to analyze the course of disability

reduction during PRP. The Pain Disability Index (PDI) was used as a measure of self-reported

disability. Many studies are aimed at the effects of PRPs and use pre- and post-treatment outcome

measures. With regard to dosage, however, the course of disability reduction during a treatment

becomes important. The duration of a program could be shortened if the desired effect could

be achieved before the end of the treatment. The analyses showed that the course of disability

reduction occurred following a quadratic model. This means that disability reduction is greater

in the beginning of the program and levels off as the treatment progresses. In addition, the

measurement moment, initial PDI score, average pain score, and interaction between the initial

PDI and measurement moment infl uenced the course of disability reduction. This means that

disability reduction is faster for a patient with higher initial PDI and average pain score than for a

patient with lower initial PDI and average pain score. The infl uence of the interaction between the

measurement moment and the initial PDI makes the reduction in PDI per measurement moment

become smaller at the end of the treatment. The results of this study can be used in clinical

practice to predict the course of PDI, and to adjust dosage to the desired effect.

R1R2R3R4R5R6R7R8R9


126 | Summary

The literature study, performed prior to the systematic review, resulted in no studies performed

to analyze dosage of PRPs as a primary objective. In Chapter 4 a randomized controlled trial is

described with regard to the use of different dosages of PRPs. The objective was to analyze the

effect of PRP with a lower dosage compared to the current program at the Center of Rehabilitation

UMCG. The treatment of the experimental group was 4 weeks shorter and had less contact hours

than that of the control group. The content of the programs was similar. The trial is of a non-

inferiority design. This means that the effect on the experimental group must not be worse than

the effect on the control group. The PDI was used as primary outcome measure. A non-inferiority

margin of 4 points difference was applied. Disability reduction was observed in both groups. The

difference between both groups was 2.5 points. This difference lies within the non-inferiority

margin. However, the 95 percent confi dence interval was -2.2 to 7.3. Because this interval is

outside the margin between 0 and 4 points, non-inferiority cannot be established for certain.

In the trial, treatments were often extended in the experimental group, making the contrast in

dosage between both groups too small. Based on a priori stopping rules, the trial was stopped

prematurely. Due to the lack of contrast in dosage between both groups and the wide margin

with regard to the precision of the results, the conclusion that a shorter treatment is non-inferior

cannot be drawn. Additionally, the difference in effects between the extended and non-extended

treatments was analyzed. The analyses did not show a signifi cant difference. This trial was the

fi rst to have the analysis of dosage as its primary goal. From the results, no robust conclusions can

be drawn for now. However, the results do indicate possibilities to shorten PRPs and the need for

further research into dosage of PRP.

Besides quantitative data regarding dosage of PRPs, it is of added value to gather qualitative

data. Chapter 5 presents a qualitative study with the aim of gaining insight into the perspectives

of patients and professionals concerning dosage of PRPs. Patients who had fi nished a PRP and

professionals of 3 rehabilitation centers in the Netherlands were interviewed, patients individually

and professionals in focus groups. The treatments of the 3 rehabilitation centers differed in

dosage. A total of 12 patients were interviewed, as well as a total of 17 professionals in 3

focus groups. Audio recordings were made of all interviews. These recordings were transcribed

verbatim, and the data was analyzed and arranged thematically. In general, patients were

satisfi ed with the dosage of treatment they had received. Although dosages differed, patients as

well as professionals mentioned similar factors regarding dosage. All factors were categorized as

patient-related, treatment-related, and external factors. Remarkably, the factors mentioned were

the same to a large extent despite the difference in dosage. Furthermore, in the focus groups

professionals reported that little is known about dosage in literature, and that is why, at the

moment, the choice of dosage of PRP is based on clinical expertise and historical grounds.


Summary | 127

S

The objective of the study described in Chapter 6 was to explore case complexity of patients

with chronic non-specifi c musculoskeletal pain (CMP). There could be a relationship between

case complexity and the required dosage of PRP. However, the case complexity of this specifi c

group was not transparent and diffi cult to measure, and that is why this relationship could not

be analyzed. Because of this, a three-round Delphi study was performed among professionals

involved in PRP in the Netherlands. These professionals were asked what they believed to be

important factors that could infl uence the functioning of patients with CMP. In the fi rst round,

a total of 166 factors were identifi ed. In the second round, the 10 most important factors were

selected. In the third round, relative weights of each of these 10 factors were calculated. A scale of

0 (no weight) to 4 (very heavy weight) was used. The 10 most important factors were: psychiatric

disorders, motivation, family system problems, mentalization capacity, somatization, fi nancial

and work related problems, features of complaints (duration, intensity), treatment-interfering

personality traits, life events, and personal injury procedures. In daily clinical practice, establishing

the infl uence of these factors on a patient’s disability is mainly based on clinical reasoning. The

10 factors were used to initiate the development of a Case Complexity Index (CCI). This index

has a minimum score of 0 and a maximum score of 100, it shows the degree of case complexity,

and is measured by establishing the presence of the 10 factors and the accompanying degree of

infl uence on case complexity. Rehabilitation physicians and professionals tested the feasibility of

the CCI on 14 patients. A reasonable similarity in CCI scores was found between the rehabilitation

physicians and professionals. This study can help to initiate the assessment of case complexity.

It may be used to gain transparency in case complexity, and to analyze the relationship between

case complexity and dosage.

In the general discussion in Chapter 7, the main fi ndings of this thesis are summarized. The factors

related to dosage of PRPs are discussed, such as factors regarding the desired effect of treatment,

patient- and PRP-related factors, and general and external factors. Clinical, scientifi c and societal

implications are described, and divided using different perspectives. Because this thesis is a fi rst

step towards scientifi c research concerning dosage of PRP, recommendations are made for future

research. The strengths and limitations of this thesis are described. The fi nal conclusion in this

chapter is that the results of this thesis offer a contribution towards: the research of PRP, making

treatments transparent and comparable, and improving cost- effectiveness.

R1R2R3R4R5R6R7R8R9


130 | Samenvatting


Samenvatting | 131

S

Pijnrevalidatie kan mensen met chronische pijn aan bewegingsapparaat helpen om beter te

leren omgaan met hun pijn en het kan de pijngerelateerde beperkingen verminderen. Deze

revalidatie bestaat uit multidisciplinaire behandelingen gebaseerd op het biopsychosociale model.

Wetenschappelijk onderzoek heeft aangetoond dat pijnrevalidatie een positief effect heeft op

beperkingen, participatie en kwaliteit van leven bij mensen met chronische pijn. De praktijkvariatie

in inhoud, vorm (klinisch - poliklinisch/ individueel - groepsbehandeling), duur en intensiteit van

de behandeling en het aantal disciplines betrokken bij de pijnrevalidatie, is echter groot. Er is nog

maar weinig bekend over de effecten van dosis van pijnrevalidatie op uitkomsten. Het doel van

de studies in dit proefschrift was dan ook het analyseren van dosisaspecten van pijnrevalidatie en

inzicht krijgen in de invloed van dosis op effecten van pijnrevalidatie. Dit is onderzocht middels

verschillende onderzoeksmethoden, beschreven in de hoofdstukken in dit proefschrift. Onder

dosis wordt in dit proefschrift verstaan: de totale duur van de behandeling in weken, het totale

aantal contacturen en het aantal contacturen per week (intensiteit).

In hoofdstuk 2 is een systematische review beschreven. De zoekstrategie in de wetenschappelijke

literatuur naar studies die primair gericht waren op het analyseren van de dosis van pijnrevalidatie,

leverde geen studies op. Daarom is gekozen om de review te richten op studies die de effecten

van pijnrevalidatie hebben geanalyseerd middels een gerandomiseerde gecontroleerde studie.

Een vereiste was dat de dosis van de onderzochte behandelingen was beschreven en het effect

gemeten was voor minimaal 1 van de uitkomsten ervaren beperkingen, arbeidsparticipatie en

kwaliteit van leven. Het resultaat is een overzicht van deze studies, de dosis van de behandelingen

en de bijbehorende omvang van effect voor de genoemde uitkomsten. De dosis van de

geïncludeerde pijnrevalidatiebehandelingen varieerde in contacturen van 6,4 tot 196,8 uur.

Aanvullend is de invloed van de dosis op het effect geanalyseerd. Een aantal variabelen bleek

invloed te hebben op het effect. Maar die variabelen waren onderling ook sterk verbonden.

Het was daarom niet mogelijk factoren te isoleren. Het was dan ook niet mogelijk om robuuste

conclusies te trekken met betrekking tot de invloed van de dosis op uitkomsten.

Hoofdstuk 3 is een prospectieve cohort studie met als doel het verloop van ervaren beperkingen,

gemeten met de Pain Disability Index (PDI), gedurende pijnrevalidatie te analyseren. Veel studies

onderzoeken het effect van pijnrevalidatie en meten uitkomstmaten vooraf en aan het eind van de

behandeling. Echter in relatie tot dosis is het verloop van de uitkomst gedurende de behandeling

interessant, omdat de behandeling verkort zou kunnen worden indien het effect al behaald is

voor het einde van de behandeling. Uit de studie bleek dat de ervaren beperkingen afnamen

volgens een kwadratisch tijd model: De mate van afname in ervaren beperkingen is in het begin

van de behandeling groter en naarmate de behandeling vordert, vlakt de afname af. Daarbij

bleek dat het meetmoment, de initiële score op de PDI, de gemiddeld pijnscore en de interactie

tussen initiële PDI en meetmoment invloed hadden op het verloop van de ervaren beperkingen.

R1R2R3R4R5R6R7R8R9


132 | Samenvatting

Klinisch betekent deze uitkomst dat de afname van ervaren beperkingen sneller gaat voor een

patiënt met een hogere initiële score op PDI en hogere pijnscore, dan voor een patiënt met een

lagere initiële score op PDI en pijnscore. De invloed van de interactie tussen het meetmoment en

de initiële PDI maakt dat de afname van PDI per meetmoment aan het einde van de behandeling

kleiner wordt. De bevindingen uit deze studie kunnen gebruikt worden om in de klinische praktijk

het verloop van de PDI te voorspellen en de dosis op het gewenste resultaat aan te passen.

Uit de literatuurstudie voorafgaand aan de systematische review, bleek dat er geen studies

waren, die primair gericht zijn op het vergelijken van pijnrevalidatie met verschillende doses. In

hoofdstuk 4 is een gerandomiseerde gecontroleerde studie beschreven waarin pijnrevalidatie

met verschillende doses is aangeboden. Het doel was het analyseren van het effect van een

pijnrevalidatiebehandeling met een lagere dosis ten opzichte van de dosis van de huidige

behandeling van het Centrum voor Revalidatie UMCG. De behandeling van de experimentele

groep was 4 weken korter en bestond uit minder contacturen dan de controle groep. De inhoud

bleef gelijkwaardig. De studie had een non-inferiority design. Dat betekent dat het effect van

de experimentele groep niet slechter mocht zijn dan het effect van de controle groep. De

PDI was gebruikt als primaire uitkomstmaat om het effect te meten, waarbij een marge van

4 punten verschil werd toegestaan als gelijkwaardig effect. Ervaren beperkingen namen af in

beide groepen. Het verschil tussen beide groepen was gemiddeld 2,5 punten. Aangezien dit

verschil binnen de marge van 4 punten ligt, wordt dit als gelijkwaardig geïnterpreteerd. De

95% betrouwbaarheidsintervallen lagen tussen -2.2 en 7.3. Aangezien de interval buiten de

gestelde marge van 0 tot 4 punten ligt, is gelijkwaardigheid niet met zekerheid vast te stellen.

In de praktijk werden de behandelingen in de experimentele groep vaak verlengd, waardoor

het contrast in dosis tussen beide groepen te klein werd. De studie moest, vanwege vooraf

opgestelde stopregels, voortijdig worden beëindigd. Het gebrek aan contrast in dosis tussen

de groepen en de grote marge in betrouwbaarheid van de resultaten, betekent dat niet met

zekerheid geconcludeerd kan worden dat het effect van een kortere behandeling gelijkwaardig

is. Aanvullend zijn analyses gedaan om het verschil tussen effect van het al dan niet verlengen van

behandelingen te analyseren. De effecten waren signifi cant verschillend. Deze studie was de eerste

waarbij dosis als primair doel werd geanalyseerd. Uit deze resultaten kunnen nog geen robuuste

conclusies getrokken worden. De resultaten geven echter wel een indicatie voor mogelijkheden

voor het verkorten van pijnrevalidatiebehandelingen en daarmee voor de noodzaak voor verder

onderzoek naar de dosis van pijnrevalidatie.

Naast kwantitatieve gegevens omtrent de dosis van pijnrevalidatie, is het ook van toegevoegde

waarde om kwalitatieve gegevens te verzamelen. Hoofdstuk 5 beschrijft een kwalitatieve studie

met als doel het verkrijgen van inzicht in de perspectieven van de patiënten en professionals

in relatie tot de dosis van pijnrevalidatie. Patiënten die een pijnrevalidatie behandeling hadden


Samenvatting | 133

S

afgerond en professionals werkzaam binnen de pijnrevalidatie in 3 revalidatiecentra in Nederland

zijn geïnterviewd. Patiënten werden individueel geïnterviewd en professionals in een groep. De

behandelingen van de 3 deelnemende revalidatiecentra verschilden in dosis. In totaal zijn 12

patiënten geïnterviewd en zijn 17 professionals in 3 groepen geïnterviewd. Van alle interviews

waren geluidsopnamen gemaakt. Deze opnamen zijn daarna letterlijk uitgeschreven en de

data zijn geanalyseerd en thematisch geordend. Over het algemeen waren patiënten tevreden

met de verkregen dosis. Ondanks de verschillen in dosis, werden zowel door de patiënten als

door de professionals soortgelijke factoren benoemd in relatie tot de dosis. Alle factoren zijn

gecategoriseerd in factoren gerelateerd aan de patiënt, aan de behandeling en aan externe

factoren. Opmerkelijk was, dat ondanks verschillen in dosis, de genoemde factoren grotendeels

hetzelfde waren tussen centra. Verder werd in de focusgroepen met de professionals aangeven,

dat er weinig bekend is over de dosis in de literatuur, waardoor de keuze van de dosis van

pijnrevalidatie op dit moment gebaseerd is op klinische ervaringen en historische gronden.

De studie beschreven in hoofdstuk 6, is opgezet om de complexiteit van de casus van de patiënten

met chronische pijnklachten die in aanmerking komen voor pijnrevalidatie, te bestuderen. Er

zou een relatie kunnen bestaan tussen de complexiteit van de casus en de benodigde dosis

van pijnrevalidatie. “Complexiteit” van de casus bleek echter voor deze specifi eke doelgroep

niet transparant en moeilijk meetbaar, waardoor de genoemde relatie niet geanalyseerd kon

worden. Vanuit de klinische praktijk leek de complexiteit te bestaan uit invloed van verschillende

factoren. In de studie namen professionals deel die werkzaam zijn binnen de pijnrevalidatie in

heel Nederland. In 3 vragenrondes werd gevraagd welke factoren een relevante invloed kunnen

hebben op het functioneren van patiënten met chronische pijnklachten. In de eerste ronde zijn

166 factoren benoemd. In de tweede ronde is hieruit een top 10 samengesteld. In de derde

ronde zijn voor de factoren uit de top 10 de mate van invloed aangegeven op het functioneren

van patiënten met chronische pijnklachten op een schaal van 0 tot 4 (0 geen invloed en 4

erg sterk van invloed). De 10 belangrijkste factoren waren: psychiatrische problematiek, lage

motivatie, systeemproblematiek, gering mentaliserend vermogen, somatisatie, fi nanciële en

werkgerelateerde problemen, klachtkenmerken (duur, intensiteit), persoonlijkheidskenmerken

die de behandeling belemmeren, life events en letselschadeprocedure. Het vaststellen van (de

mate van invloed van) deze factoren bleek in de huidige praktijk vooral gebaseerd op klinisch

redeneren; niet op basis van uitkomsten van meetinstrumenten. De 10 factoren zijn gebruikt om

een casus complexiteitsindex (CCI) op te stellen. Deze index, geeft de mate van complexiteit weer.

De CCI wordt berekend op basis van de aanwezigheid van de 10 factoren en de bijbehorende

mate van invloed op complexiteit. De score kan variëren van 0 tot 100, waarbij een hoge score

duidt op een hoge casus complexiteit. De score kan bestaan uit aanwezigheid van meerdere

factoren of een paar factoren met een hoge mate van invloed. Het gebruik van de CCI is getest

in de huidige praktijk door revalidatieartsen en het behandelteam bij 14 patiënten. Er bleek een

R1R2R3R4R5R6R7R8R9


134 | Samenvatting

redelijke overeenstemming van score van CCI tussen de revalidatieartsen en het behandelteam.

Deze studie heeft een eerste aanzet gedaan naar het meetbaar maken van de complexiteit van

de casus. De resultaten van deze studie kunnen mogelijk gebruikt worden om transparantie te

krijgen in casuscomplexiteit en om de relatie tussen complexiteit en dosis te analyseren.

In de algemene discussie in hoofdstuk 7 zijn de voornaamste resultaten van het proefschrift

beschreven. De verschillende factoren die gerelateerd zijn aan de dosis van PRP, zoals factoren

met betrekking tot het beoogde behandeleffect, patiënt en PRP gerelateerde factoren, algemene

en externe factoren, worden benoemd en bediscussieerd. De klinische, wetenschappelijke en

maatschappelijke implicaties worden vanuit verschillende perspectieven beschreven. Omdat

dit proefschrift een eerste stap is geweest om de dosis van pijnrevalidatie wetenschappelijk te

onderzoeken, zijn aanbevelingen gedaan voor vervolgonderzoek. De sterke en minder sterke

punten van het proefschrift zijn beschreven. Het hoofdstuk wordt afgesloten met de conclusie dat

de resultaten van dit proefschrift kunnen bijdragen aan onderzoek binnen de pijnrevalidatie, het

transparant en vergelijkbaar maken van keuzes ten aanzien van de dosis en op termijn mogelijk

tot het verbeteren van de kosteneffectiviteit.


Samenvatting | 135

S

R1R2R3R4R5R6R7R8R9


138 | Dankwoord


Dankwoord | 139

D

In september 2009 ben ik gestart met mijn promotietraject. Een enorme uitdaging; dat was het

zeker!!! Gelukkig heb ik vaak steun ervaren van de mensen om mij heen, met wie ik vele goede,

en zeker ook de minder goede momenten heb mogen delen. Via deze weg wil ik iedereen

bedanken en een aantal mensen in het bijzonder:

Allereerst alle professionals en patiënten die deel hebben genomen aan de verschillende

onderzoeken. Zonder deelnemers, geen onderzoek. Dank voor ieders inzet en bijdrage.

Dank aan mijn promotieteam: prof. dr. JHB Geertzen, prof. dr. MF Reneman en prof. dr. PU

Dijkstra. Voor mij was dit promotieteam de perfecte combinatie. Balans tussen actie en rust;

inhoud en praktijk en de zuidelijke gezelligheid en de noordelijke stugheid. Ik dank jullie voor de

prettige samenwerking en alles wat ik van jullie heb geleerd.

Beste Jan, jij gaf mij de kans om als ergotherapeut te promoveren. De eerste jaren was jij mijn

eerste promotor. Zodra Michiel benoemd werd tot professor, werd hij eerste promotor. Stiekem

ben je, naast Michiel, ook altijd een beetje mijn eerste promotor gebleven! Dank voor al de

kritische noten, de waardevolle aanvullingen en natuurlijk de Brabantse touch!

Beste Michiel, ik weet nog dat je kort na de promotie van Remko tegen mij zei: “Wij hadden

bedacht dat jij over een aantal jaren ook wel in die aula kon gaan staan”. Het duurde bij mij

even voordat ik in de gaten had waar je het over had. Dank dat je mij deze kans hebt gegeven

en dank voor de vele steun, het inkorten en afkappen van mijn ellenlange zinnen en discussies.

Maar bovenal voor het vertrouwen dat je me gegeven hebt. Als ik even door de bomen het bos

niet meer zag, kon je me in laten zien hoe mooi het project was en wat het onderzoek allemaal

opleverde! Daardoor raakte ik, ondanks de obstakels die ik moest nemen, toch weer gemotiveerd

om positief verder te gaan.

Beste Pieter, jij kwam iets later bij de club. Wat ben ik vaak blij geweest met jouw expertise. Al die

data en al die statistische formules waar ik niets van snapte en wat jij met alle geduld uitlegde.

Samen zochten we naar de vertaalslag van de resultaten naar de betekenis voor klinische praktijk.

Dank voor je geduld, je wijze raad en de goede en gezellige gesprekken.

Het nemen van de stap om als ergotherapeut wetenschappelijk onderzoek te gaan doen en

uiteindelijk te gaan promoveren, heb ik voornamelijk te danken aan Nelleke Kombrink, Rita

Schiphorst Preuper en Michiel Reneman. Zij hebben hier een belangrijke bijdrage aan geleverd.

Nelleke, jij bent degene die mij heeft aangespoord om de wetenschappelijke kant op te gaan. Jij was

teammanager van het pijnteam en tijdens mijn jaargesprek kwamen de onderwerpen uitdaging

en wetenschap bij elkaar. Je stimuleerde me om mijzelf te oriënteren op de mogelijkheden een

wetenschappelijke studie te gaan doen. Ik ben je dankbaar voor het advies wat je me toen

gegeven hebt met dit proefschrift als resultaat!

R1R2R3R4R5R6R7R8R9


140 | Dankwoord

Rita, jij stimuleerde dit ook. Je dacht mee over de mogelijkheden om te kunnen promoveren en

adviseerde mij tijdens het hele promotietraject. Jouw klinische en wetenschappelijke blik en de

daaruit voortkomende zinvolle aanvullingen en ideeën hebben de uitvoer van het hele project

goed gedaan! Dank hiervoor.

Annemieke de Jong en Berry van Holland, wat geweldig dat jullie mijn paranimfen willen zijn!

Annemieke: als collega’s in het pijnteam, als onderzoekers en zeker niet op de laatste plaats ook

als dames die graag samen een hapje gaan eten om bij te kletsen, hebben we vele gesprekken

gehad en heb ik veel aan je gehad. Dank voor alles: voor de gesprekken, adviezen en gezellige

eetafspraken!

Berry, je bent de beste kamergenoot die iemand zich kan wensen! Of het nu ging om de goede

Engelse bewoording, de lastige formules in Excel, het vieren van de successen of het uiten van

de frustraties, je had altijd een luisterend oor en was altijd behulpzaam! Heel, heel erg bedankt

daarvoor!

Alle collega’s en ex-collega’s van het pijnteam wil ik bedanken voor alle steun, acties, bijdragen,

kritische blikken, discussies, interesses en ieders luisterend oor. De revalidatieartsen Maikel, Jitze,

Rita en Wim: speciale dank voor het informeren en includeren van alle patiënten voor de RCT en

ieders bijdrage aan het onderzoek en dit proefschrift. Alle ergotherapeuten, fysiotherapeuten,

psychologen en maatschappelijk werkster bedankt! Zonder jullie inzet was het niet gelukt en

zonder jullie steun was het mij nooit gelukt om dit tot een goed einde te brengen. Speciale dank

voor mijn directe collega’s van de ergotherapie: Anne, Judith en Lucienne. Ik kon altijd bij jullie

terecht, dat waardeer ik enorm.

De secretaresses van het pijnteam en in het bijzonder Marleen Speller: dank voor alle hulp die

jullie geboden hebben gedurende het hele traject en met name tijdens de uitvoer van de RCT.

Wat kunnen jullie veel werk verzetten en ook nog overzicht houden! Dank hiervoor.

Alle coauteurs wil ik bedanken voor hun waardevolle bijdrage aan het tot stand komen van de

verschillende artikelen in dit proefschrift.

Tijdens een promotietraject kruisen vele praktische en emotionele zaken je pad. Wat was ik dan

ook altijd blij met de collega’s die dezelfde ervaringen hadden op het gebied van onderzoek doen

en promoveren. De collega-onderzoekers van het pijnteam: Remko, Haitze, Berry en de collega’s

van Oker: Sacha, Ilse, Feyuna, Christa, Ria, Anja en Ant. Dank voor alle adviezen en het delen van

jullie (persoonlijke) ervaringen en expertise.


Dankwoord | 141

D

Tenslotte wil ik mijn lieve familie en vrienden bedanken voor alle steun en belangstelling.

Er was altijd tijd en belangstelling voor mijn project.

Lieve pap, mam en Ellen, dank dat jullie er altijd voor me zijn!

Lieve Gino, jij hebt altijd geloofd en gezegd dat ik dit kon, ondanks dat ik daar zelf nog niet zo

overtuigd van was. En je had gelijk: het is gelukt! Thanx laiverd!

R1R2R3R4R5R6R7R8R9


144 | Over de auteur


Over de auteur | 145

A

Franka Waterschoot, geboren op 20 juni 1978 te Hoogeloon, groeide op in de buurt van

Eindhoven. Na de middelbare school is ze de opleiding tot ergotherapeut gaan volgen in Heerlen,

waar ze in 2002 afstudeerde. Nadien verhuisde ze van Noord Brabant naar Groningen. Het

eerste half jaar heeft ze gewerkt als ergotherapeut in verpleegtehuis Dilgtoord in Haren en de

daarbij horende locatie de Burcht in Hoogezand. Vanaf oktober 2002 is zij gaan werken bij het

UMCG, Centrum voor Revalidatie (CvR), locatie Beatrixoord Haren. Binnen Beatrixoord heeft ze

als ergotherapeut met name gewerkt binnen de pijnrevalidatie, maar ook binnen de diabetes-

en neurorevalidatie. Na een aantal jaren was Franka op zoek naar een nieuwe uitdaging. Deze

uitdaging startte in september 2009 met een parttime promotietraject met dit proefschrift als

eindresultaat. Sinds oktober 2014 heeft Franka de shift gemaakt van de patiëntenzorg naar

de kwaliteit van de zorg en zet ze haar kennis en expertise in voor projecten binnen het CvR.

Momenteel is ze projectleider van de implementatie van de landelijke behandelmodules binnen

het CvR en van de werkgroep die zich bezig houdt met de zorglogistiek ten behoeve van een

gastvrij en patiëntgericht CvR. Daarnaast is ze coördinator van innovaties in relatie tot ehealth en

zorglogistieke processen binnen het pijnteam.

R1R2R3R4R5R6R7R8R9


148 | Research Institute SHARE


Research Institute SHARE | 149

R

Research Institute SHARE

This thesis is published within the Research Institute SHARE (Science in Healthy

Ageing and healthcaRE) of the University Medical Center Groningen / University of

Groningen.

Further information regarding the institute and its research can be obtained from our internetsite:

http://www.share.umcg.nl/

More recent theses can be found in the list below.

((co-) supervisors are between brackets)

Janse M

The art of adjustment

(prof AV Ranchor, prof MAG Sprangers, dr J Fleer)

Nigatu YT

Obesity and depression; an intertwined public health challenge

(prof U Bültmann, prof SA Reijneveld)

Aris-Meijer JL

Stormy clouds in seventh heaven; a study on anxiety and depression around

childbirth

(prof CLH Bockting, prof RP Stolk, dr H Burger)

Kluitenberg B

The NLstart2run study; running related injuries in novice runners

(prof RL Diercks, dr H van der Worp, dr M van Middelkoop)

Benjaminse A

Motor learning in ACL injury prevention

(Prof E Otten, prof KAPM Lemmink, prof RL Diecks)

Potijk MR

Moderate prematurity, socioeconomic status, and neurodevelopment in early

childhood; a life course perspective

(prof SA Reijneveld, prof AF Bos, dr AF de Winter)

R1R2R3R4R5R6R7R8R9


150 | Research Institute SHARE

Beyer AR

The human dimension in the assessment of medicines; perception, preferences,

and decision making in the European regulatory environment

(prof JL Hillege, prof PA de Graeff, prof B Fasolo)

Alferink M

Psychological factors related to Buruli ulcer and tuberculosis in Sub-saharan Africa

(prof AV Ranchor, prof TS van der Werf, dr Y Stienstra)

Bennik EC

Every dark cloud has a colored lining; the relation between positive and negative

affect and reactivity to positive and negative events

(prof AJ Oldehinkel, prof J Ormel, dr E Nederhof, dr JACJ Bastiaansen)

Stallinga HA

Human functioning in health care; application of the International Classifi cation of

Functioning, Disability and Health (ICF)

(prof RF Roodbol, prof PU Dijkstra, dr G Jansen)

Papageorgiou A

How is depression valued?

(prof AV Ranchor, prof E Buskens, dr KM Vermeulen, dr MJ Schroevers)

Spijkers W

Parenting and child psychosocial problems; effectiveness of parenting support in

preventive child healthcare

(prof SA Reijneveld, dr DEMC Jansen)

Ark M van

Patellar tendinopathy; physical therapy and injection treatments

(prof RL Diercks, prof JL Cook)

Boven JFM van

Enhancing adherence in patients with COPD: targets, interventions and cost-

effectiveness

(prof MJ Postma, prof T van der Molen)


Research Institute SHARE | 151

R

Gerbers JG

Computer assisted surgery in orthopedaedic oncology; indications, applications

and surgical workfl ow

(prof SK Bulstra, dr PC Jutte, dr M Stevens)

Niet A van der

Physical activity and cognition in children

(prof C Visscher, dr E Hartman, dr J Smith)

For more 2015 and earlier theses visit our website

R1R2R3R4R5R6R7R8R9


154 | EXPAND


EXPAND | 155

E

Wetenschappelijk onderzoek afdeling Revalidatiegeneeskunde –Centrum voor Revalidatie UMCG

EXPANDExtremities, Pain and Disability

Missie: EXPAND draagt bij aan participatie en kwaliteit van leven van mensen met aandoeningen en

amputaties van de extremiteiten of met pijn aan het bewegingsapparaat.

EXPAND omvat twee speerpunten:• onderzoek naar aandoeningen aan en amputaties van extremiteiten met nadruk op

stoornissen, activiteiten en participatie,• onderzoek naar chronische pijn en arbeidsparticipatie.

EXPAND draagt bij aan het UMCG-brede thema Healthy Ageing.

Research Department of Rehabilitation Medicine –Center for Rehabilitation UMCG

EXPANDExtremities, Pain and Disability

Mission:EXPAND contributes to participation and quality of life of people with health conditions and

amputations of the extremities and musculoskeletal pain.

EXPAND includes two spearheads:• research on the health conditions and amputations of the extremities with emphasis on

body functions and structures, activities and participations,• chronic pain and work participation.

EXPAND contributes to Healthy Aging, the focus of the UMCG.